config.json

{
  "architectures": [
    "LlavaNextForConditionalGeneration"
  ],
  "dtype": "bfloat16",
  "image_grid_pinpoints": [
    [
      112,
      112
    ],
    [
      112,
      224
    ],
    [
      224,
      112
    ]
  ],
  "image_seq_length": 576,
  "image_token_index": 49155,
  "model_type": "llava_next",
  "multimodal_projector_bias": true,
  "projector_hidden_act": "gelu",
  "text_config": {
    "architectures": [
      "GraniteForCausalLM"
    ],
    "attention_bias": false,
    "attention_dropout": 0.1,
    "attention_multiplier": 0.015625,
    "bos_token_id": 0,
    "dtype": "bfloat16",
    "embedding_multiplier": 12.0,
    "eos_token_id": 0,
    "head_dim": 32,
    "hidden_act": "silu",
    "hidden_size": 64,
    "initializer_range": 0.02,
    "intermediate_size": 128,
    "logits_scaling": 8.0,
    "max_position_embeddings": 131072,
    "mlp_bias": false,
    "model_type": "granite",
    "num_attention_heads": 2,
    "num_hidden_layers": 2,
    "num_key_value_heads": 2,
    "pad_token_id": 0,
    "residual_multiplier": 0.22,
    "rms_norm_eps": 1e-05,
    "rope_parameters": {
      "rope_theta": 300000,
      "rope_type": "default"
    },
    "tie_word_embeddings": true,
    "use_cache": true,
    "vocab_size": 49156
  },
  "tie_word_embeddings": true,
  "transformers_version": "5.8.1",
  "use_image_newline_parameter": true,
  "vision_config": {
    "attention_dropout": 0.0,
    "hidden_act": "gelu_pytorch_tanh",
    "hidden_size": 64,
    "image_size": 112,
    "intermediate_size": 128,
    "layer_norm_eps": 1e-06,
    "model_type": "siglip_vision_model",
    "num_attention_heads": 4,
    "num_channels": 3,
    "num_hidden_layers": 4,
    "patch_size": 14
  },
  "vision_feature_layer": [
    -4,
    -3,
    -2,
    -1
  ],
  "vision_feature_select_strategy": "full"
}