{ "_flash_attn_2_enabled": true, "architectures": [ "LlamaForCausalLM" ], "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 128000, "dtype": "float32", "eos_token_id": 128001, "head_dim": 64, "hidden_act": "silu", "hidden_size": 1536, "initializer_range": 0.02, "intermediate_size": 4224, "max_position_embeddings": 3072, "mlp_bias": false, "model_type": "llama", "num_attention_heads": 24, "num_hidden_layers": 22, "num_key_value_heads": 8, "pad_token_id": 128001, "pretraining_tp": 1, "rms_norm_eps": 1e-05, "rope_scaling": null, "rope_theta": 10000.0, "tie_word_embeddings": false, "transformers_version": "4.56.2", "use_cache": false, "vocab_size": 128256 }