| { | |
| "_name_or_path": "/home/s9dazhuu/outputs/exp_outputs/alpaca_mt/alpaca_mt_one_arc_sample-mt-true_random-en-cs-wmt17-wmt18-wmt19-wmt20-dN-1-dseed42-ep3-lr2e5-bs4-accu8-seed42/trained_models", | |
| "architectures": [ | |
| "LlamaForCausalLM" | |
| ], | |
| "bos_token_id": 1, | |
| "eos_token_id": 2, | |
| "hidden_act": "silu", | |
| "hidden_size": 4096, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 11008, | |
| "max_position_embeddings": 4096, | |
| "model_type": "llama", | |
| "num_attention_heads": 32, | |
| "num_hidden_layers": 32, | |
| "num_key_value_heads": 32, | |
| "pretraining_tp": 1, | |
| "rms_norm_eps": 1e-05, | |
| "rope_scaling": null, | |
| "tie_word_embeddings": false, | |
| "torch_dtype": "float32", | |
| "transformers_version": "4.32.1", | |
| "use_cache": true, | |
| "vocab_size": 32001 | |
| } | |