| { |
| "architectures": [ |
| "VoxtralForConditionalGeneration" |
| ], |
| "audio_config": { |
| "activation_dropout": 0.0, |
| "activation_function": "gelu", |
| "attention_dropout": 0.0, |
| "dropout": 0.0, |
| "head_dim": 32, |
| "hidden_size": 64, |
| "initializer_range": 0.02, |
| "intermediate_size": 256, |
| "layerdrop": 0.0, |
| "max_source_positions": 1500, |
| "model_type": "voxtral_encoder", |
| "num_attention_heads": 2, |
| "num_hidden_layers": 2, |
| "num_key_value_heads": 2, |
| "num_mel_bins": 128, |
| "scale_embedding": false, |
| "vocab_size": 51866 |
| }, |
| "audio_token_id": 24, |
| "hidden_size": 64, |
| "model_type": "voxtral", |
| "projector_hidden_act": "gelu", |
| "text_config": { |
| "attention_bias": false, |
| "attention_dropout": 0.0, |
| "head_dim": 32, |
| "hidden_act": "silu", |
| "hidden_size": 64, |
| "initializer_range": 0.02, |
| "intermediate_size": 128, |
| "max_position_embeddings": 131072, |
| "mlp_bias": false, |
| "model_type": "llama", |
| "num_attention_heads": 2, |
| "num_hidden_layers": 2, |
| "num_key_value_heads": 1, |
| "pretraining_tp": 1, |
| "rms_norm_eps": 1e-05, |
| "rope_scaling": null, |
| "rope_theta": 100000000.0, |
| "sliding_window": null, |
| "tie_word_embeddings": true, |
| "use_cache": true, |
| "vocab_size": 131072 |
| }, |
| "torch_dtype": "bfloat16", |
| "transformers_version": "4.54.0.dev0", |
| "vocab_size": 131072 |
| } |
|
|
