{ "architectures": [ "Qwen2ForCausalLM" ], "hidden_size": 2560, "intermediate_size": 6144, "max_position_embeddings": 1024, "model_type": "qwen2", "num_attention_heads": 40, "num_hidden_layers": 30, "qkv_bias": true, "torch_dtype": "float32", "transformers_version": "4.47.1", "vocab_size": 151936 }