| { |
| "architectures": [ |
| "NemotronHAudexForConditionalGeneration" |
| ], |
| "attention_bias": false, |
| "attention_dropout": 0.0, |
| "audio_config": { |
| "activation_function": "gelu", |
| "d_model": 1280, |
| "encoder_attention_heads": 20, |
| "encoder_ffn_dim": 5120, |
| "encoder_layers": 32, |
| "max_source_positions": 1500, |
| "model_type": "qwen2_audio_encoder", |
| "num_mel_bins": 128, |
| "scale_embedding": false |
| }, |
| "audio_encoder_hidden_size": 1280, |
| "audio_model_type": "NV-Whisper", |
| "audio_preprocessor_path": "audio_preprocessor", |
| "audio_projector_activation": "relu2", |
| "audio_projector_intermediate_size": 4096, |
| "audio_projector_norm_eps": 1e-05, |
| "auto_map": { |
| "AutoConfig": "configuration_nemotron_h_audio.NemotronHAudexConfig", |
| "AutoModel": "modeling_nemotron_h_audio.NemotronHAudexForConditionalGeneration", |
| "AutoModelForCausalLM": "modeling_nemotron_h_audio.NemotronHAudexForConditionalGeneration" |
| }, |
| "base_model_tp_plan": null, |
| "bos_token_id": 1, |
| "chunk_size": 128, |
| "conv_kernel": 4, |
| "eos_token_id": 11, |
| "expand": 2, |
| "head_dim": 128, |
| "hidden_dropout": 0.0, |
| "hidden_size": 2688, |
| "hybrid_override_pattern": "MEMEM*EMEMEM*EMEMEM*EMEMEM*EMEMEM*EMEMEMEM*EMEMEMEME", |
| "initializer_range": 0.02, |
| "intermediate_size": 1856, |
| "layer_norm_epsilon": 1e-05, |
| "mamba_head_dim": 64, |
| "mamba_hidden_act": "silu", |
| "mamba_num_heads": 64, |
| "mamba_proj_bias": false, |
| "max_position_embeddings": 262144, |
| "mlp_bias": false, |
| "mlp_hidden_act": "relu2", |
| "model_type": "nemotron_h_audex", |
| "moe_intermediate_size": 1856, |
| "moe_shared_expert_intermediate_size": 3712, |
| "n_group": 1, |
| "n_groups": 8, |
| "n_routed_experts": 128, |
| "n_shared_experts": 1, |
| "norm_eps": 1e-05, |
| "norm_topk_prob": true, |
| "num_attention_heads": 32, |
| "num_experts_per_tok": 6, |
| "num_hidden_layers": 52, |
| "num_key_value_heads": 2, |
| "num_logits_to_keep": 1, |
| "pad_token_id": 0, |
| "partial_rotary_factor": 1.0, |
| "rescale_prenorm_residual": true, |
| "residual_in_fp32": false, |
| "rope_theta": 10000, |
| "routed_scaling_factor": 2.5, |
| "sliding_window": null, |
| "sound_clip_duration": 30.0, |
| "sound_embedding_size": 750, |
| "sound_end_token": "<so_end>", |
| "sound_end_token_id": 31, |
| "sound_model_type": "hf:///nv-whisper", |
| "sound_start_token": "<so_start>", |
| "sound_start_token_id": 30, |
| "sound_target_rate": 16000, |
| "sound_token": "<so_embedding>", |
| "sound_token_id": 29, |
| "ssm_state_size": 128, |
| "tie_word_embeddings": false, |
| "time_step_floor": 0.0001, |
| "time_step_max": 0.1, |
| "time_step_min": 0.001, |
| "topk_group": 1, |
| "tp_plan": null, |
| "torch_dtype": "bfloat16", |
| "dtype": "bfloat16", |
| "transformers_version": "4.53.0", |
| "use_bias": false, |
| "use_cache": true, |
| "use_conv_bias": true, |
| "use_mamba_kernels": true, |
| "vocab_size": 205312, |
| "quantization": { |
| "group_size": 64, |
| "bits": 4, |
| "mode": "affine" |
| }, |
| "weight_format": "affine" |
| } |
|
|
