{ "_name_or_path": "MiniCPM5-1B-MLX", "architectures": [ "LlamaForCausalLM" ], "bos_token_id": 0, "eos_token_id": [ 1, 130073 ], "head_dim": 128, "hidden_act": "silu", "hidden_size": 1536, "initializer_range": 0.02, "intermediate_size": 4608, "max_position_embeddings": 131072, "model_type": "llama", "num_attention_heads": 16, "num_hidden_layers": 24, "num_key_value_heads": 2, "pad_token_id": 1, "quantization": { "group_size": 64, "bits": 4, "mode": "affine" }, "quantization_config": { "group_size": 64, "bits": 4, "mode": "affine" }, "rms_norm_eps": 1e-06, "rope_scaling": null, "rope_theta": 5000000, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "transformers_version": "5.6.2", "use_cache": true, "vocab_size": 130560 }