{ "_name_or_path": "/tmp/tmpa7qdhm_p", "architectures": [ "LlamaForCausalLM" ], "attention_bias": true, "attention_dropout": 0.1, "bos_token_id": 0, "eos_token_id": 0, "hidden_act": "silu", "hidden_size": 2560, "initializer_range": 0.02, "intermediate_size": 10240, "max_position_embeddings": 2048, "mlp_bias": true, "model_type": "llama", "num_attention_heads": 32, "num_hidden_layers": 32, "num_key_value_heads": 32, "pad_token_id": 0, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 10000, "tie_word_embeddings": true, "torch_dtype": "float16", "transformers_version": "4.42.4", "use_cache": true, "vocab_size": 49152 }