| { | |
| "architectures": [ | |
| "LlamaForCausalLM" | |
| ], | |
| "attention_bias": false, | |
| "attention_dropout": 0.0, | |
| "bos_token_id": 48, | |
| "dtype": "bfloat16", | |
| "eos_token_id": 48, | |
| "head_dim": 128, | |
| "hidden_act": "silu", | |
| "hidden_size": 2816, | |
| "initializer_range": 0.005, | |
| "intermediate_size": 9856, | |
| "max_position_embeddings": 65536, | |
| "mlp_bias": false, | |
| "model_type": "llama", | |
| "num_attention_heads": 48, | |
| "num_hidden_layers": 60, | |
| "num_key_value_heads": 8, | |
| "pad_token_id": 48, | |
| "pretraining_tp": 1, | |
| "rms_norm_eps": 3e-06, | |
| "rope_parameters": { | |
| "beta_fast": 32.0, | |
| "beta_slow": 1.0, | |
| "factor": 10.0, | |
| "original_max_position_embeddings": 8192, | |
| "rope_theta": 200000, | |
| "rope_type": "yarn", | |
| "truncate": true, | |
| "type": "yarn" | |
| }, | |
| "tie_word_embeddings": false, | |
| "transformers_version": "5.3.0", | |
| "use_cache": false, | |
| "vocab_size": 131072 | |
| } | |