{ "_name_or_path": "/cpfs/user/chennuo/CN/output/deepspeed/xlingual/output_step1_llama_33b_mixV1_2e-5_bf16/", "architectures": [ "LlamaModel" ], "bos_token_id": 1, "end_token_id": 2, "eos_token_id": 2, "hidden_act": "silu", "hidden_size": 6656, "initializer_range": 0.02, "intermediate_size": 17920, "max_position_embeddings": 2048, "model_type": "llama", "num_attention_heads": 52, "num_hidden_layers": 60, "pad_token_id": 2, "rms_norm_eps": 1e-06, "tie_word_embeddings": false, "torch_dtype": "float32", "transformers_version": "4.29.2", "use_cache": true, "vocab_size": 32000 }