| { | |
| "architectures": [ | |
| "GPTNeoXForCausalLM" | |
| ], | |
| "bos_token_id": 1, | |
| "eos_token_id": 1, | |
| "hidden_act": "gelu", | |
| "hidden_size": 2560, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 10240, | |
| "layer_norm_eps": 1e-05, | |
| "max_position_embeddings": 8192, | |
| "model_type": "gpt_neox", | |
| "num_attention_heads": 20, | |
| "num_hidden_layers": 32, | |
| "rotary_emb_base": 10000, | |
| "rotary_pct": 1.0, | |
| "tie_word_embeddings": false, | |
| "torch_dtype": "float16", | |
| "transformers_version": "4.24.0", | |
| "use_cache": true, | |
| "use_parallel_residual": false, | |
| "vocab_size": 64000 | |
| } | |