{ "_name_or_path": "/Users/guyuxian/miniplm/200M/0", "architectures": [ "Qwen2ForCausalLM" ], "attention_dropout": 0.0, "bos_token_id": 151643, "eos_token_id": 151643, "hidden_act": "silu", "hidden_size": 768, "initializer_range": 0.02, "intermediate_size": 2112, "is_model_parallel": false, "max_position_embeddings": 32768, "max_window_layers": 10, "model_type": "qwen2", "num_attention_heads": 12, "num_hidden_layers": 12, "num_key_value_heads": 12, "rms_norm_eps": 1e-06, "rope_scaling": null, "rope_theta": 1000000.0, "sliding_window": null, "tie_word_embeddings": true, "torch_dtype": "float32", "transformers_version": "4.45.2", "use_cache": true, "use_memory_efficient_attention": true, "use_sliding_window": false, "vocab_size": 151936 }