{ "_name_or_path": "/workspace/ultravox_/runs/exp--2025-05-31--17-00-24/checkpoint-125", "architectures": [ "UltravoxModel" ], "audio_latency_block_size": null, "audio_model_id": "openai/whisper-large-v3-turbo", "auto_map": { "AutoConfig": "ultravox_config.UltravoxConfig", "AutoModel": "ultravox_model.UltravoxModel" }, "custom_pipelines": { "ultravox-pipeline": { "impl": "ultravox_pipeline.UltravoxPipeline", "pt": [ "AutoModel" ], "tf": [], "type": "multimodal" } }, "hidden_size": 4096, "ignore_index": -100, "initializer_range": 0.02, "model_type": "ultravox", "norm_init": 0.4, "pad_token_id": 128009, "projector_act": "swiglu", "projector_ln_mid": true, "stack_factor": 8, "text_config": { "_name_or_path": "meta-llama/Llama-3.2-1B-Instruct", "architectures": [ "LlamaForCausalLM" ], "bos_token_id": 128000, "eos_token_id": [ 128001, 128008, 128009 ], "head_dim": 64, "hidden_size": 2048, "intermediate_size": 8192, "max_position_embeddings": 131072, "model_type": "llama", "num_hidden_layers": 16, "num_key_value_heads": 8, "rms_norm_eps": 1e-05, "rope_scaling": { "factor": 32.0, "high_freq_factor": 4.0, "low_freq_factor": 1.0, "original_max_position_embeddings": 8192, "rope_type": "llama3" }, "rope_theta": 500000.0, "tie_word_embeddings": true, "torch_dtype": "bfloat16", "vocab_size": 128256 }, "text_model_id": null, "torch_dtype": "bfloat16", "transformers_version": "4.48.1", "vocab_size": 128256 }