diff --git "a/config.json" "b/config.json"
new file mode 100644--- /dev/null
+++ "b/config.json"
@@ -0,0 +1,9449 @@
+{
+    "architectures": [
+        "GlmMoeDsaForCausalLM"
+    ],
+    "attention_bias": false,
+    "attention_dropout": 0.0,
+    "dtype": "bfloat16",
+    "eos_token_id": [
+        154820,
+        154827,
+        154829
+    ],
+    "ep_size": 1,
+    "first_k_dense_replace": 3,
+    "head_dim": 64,
+    "hidden_act": "silu",
+    "hidden_size": 6144,
+    "index_head_dim": 128,
+    "index_n_heads": 32,
+    "index_topk": 2048,
+    "indexer_rope_interleave": true,
+    "initializer_range": 0.02,
+    "intermediate_size": 12288,
+    "kv_lora_rank": 512,
+    "max_position_embeddings": 202752,
+    "model_type": "glm_moe_dsa",
+    "moe_intermediate_size": 2048,
+    "moe_layer_freq": 1,
+    "n_group": 1,
+    "n_routed_experts": 256,
+    "n_shared_experts": 1,
+    "norm_topk_prob": true,
+    "num_attention_heads": 64,
+    "num_experts_per_tok": 8,
+    "num_hidden_layers": 78,
+    "num_key_value_heads": 64,
+    "num_nextn_predict_layers": 1,
+    "pad_token_id": 154820,
+    "pretraining_tp": 1,
+    "q_lora_rank": 2048,
+    "qk_head_dim": 256,
+    "qk_nope_head_dim": 192,
+    "qk_rope_head_dim": 64,
+    "rms_norm_eps": 1e-05,
+    "rope_interleave": true,
+    "rope_parameters": {
+        "rope_theta": 1000000,
+        "rope_type": "default"
+    },
+    "routed_scaling_factor": 2.5,
+    "scoring_func": "sigmoid",
+    "tie_word_embeddings": false,
+    "topk_group": 1,
+    "topk_method": "noaux_tc",
+    "transformers_version": "5.0.2.dev0",
+    "use_cache": true,
+    "v_head_dim": 256,
+    "vocab_size": 154880,
+    "quantization": {
+        "group_size": 32,
+        "bits": 4,
+        "mode": "mxfp4",
+        "model.embed_tokens": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "lm_head": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.mlp.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.mlp.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.mlp.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.mlp.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.mlp.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.mlp.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.mlp.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.mlp.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.mlp.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        }
+    },
+    "quantization_config": {
+        "group_size": 32,
+        "bits": 4,
+        "mode": "mxfp4",
+        "model.embed_tokens": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "lm_head": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.mlp.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.mlp.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.0.mlp.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.mlp.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.mlp.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.1.mlp.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.mlp.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.mlp.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.2.mlp.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.3.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.4.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.5.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.6.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.7.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.8.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.9.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.10.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.11.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.12.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.13.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.14.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.15.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.16.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.17.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.18.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.19.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.20.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.21.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.22.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.23.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.24.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.25.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.26.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.27.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.28.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.29.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.30.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.31.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.32.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.33.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.34.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.35.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.36.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.37.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.38.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.39.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.40.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.41.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.42.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.43.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.44.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.45.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.46.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.47.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.48.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.49.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.50.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.51.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.52.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.53.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.54.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.55.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.56.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.57.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.58.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.59.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.60.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.61.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.62.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.63.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.64.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.65.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.66.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.67.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.68.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.69.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.70.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.71.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.72.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.73.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.74.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.75.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.76.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.self_attn.q_a_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.self_attn.q_b_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.self_attn.kv_a_proj_with_mqa": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.self_attn.embed_q": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.self_attn.unembed_out": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.self_attn.o_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.self_attn.indexer.wq_b": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.self_attn.indexer.wk": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.self_attn.indexer.weights_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.mlp.shared_experts.gate_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.mlp.shared_experts.up_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        },
+        "model.layers.77.mlp.shared_experts.down_proj": {
+            "group_size": 64,
+            "bits": 8,
+            "mode": "affine"
+        }
+    }
+}
\ No newline at end of file