| { | |
| "architectures": [ | |
| "LlamaForCausalLM" | |
| ], | |
| "attention_bias": false, | |
| "attention_dropout": 0.0, | |
| "attention_probs_dropout_prob": 0.1, | |
| "bos_token_id": 0, | |
| "dtype": "bfloat16", | |
| "eos_token_id": 2, | |
| "head_dim": 128, | |
| "hidden_act": "silu", | |
| "hidden_dropout_prob": 0.0, | |
| "hidden_size": 1792, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 4608, | |
| "max_position_embeddings": 32768, | |
| "mlp_bias": false, | |
| "model_type": "llama", | |
| "num_attention_heads": 32, | |
| "num_hidden_layers": 48, | |
| "num_key_value_heads": 8, | |
| "pretraining_tp": 1, | |
| "quantization_config": { | |
| "modules_to_not_convert": [ | |
| "lm_head", | |
| "model.embed_tokens" | |
| ], | |
| "quant_method": "mxfp4" | |
| }, | |
| "rms_norm_eps": 1e-05, | |
| "rope_scaling": null, | |
| "rope_theta": 8000000, | |
| "tie_word_embeddings": true, | |
| "transformers_version": "4.57.1", | |
| "use_cache": false, | |
| "vocab_size": 131392 | |
| } | |