| { | |
| "architectures": [ | |
| "GPTNeoXJapaneseForCausalLM" | |
| ], | |
| "model_type": "gpt_neox_japanese", | |
| "vocab_size": 32000, | |
| "hidden_size": 2560, | |
| "num_hidden_layers": 32, | |
| "num_attention_heads": 32, | |
| "intermediate_multiple_size": 4, | |
| "hidden_act": "gelu", | |
| "rotary_pct": 1.00, | |
| "rotary_emb_base": 10000, | |
| "max_position_embeddings": 2048, | |
| "initializer_range": 0.02, | |
| "layer_norm_eps": 1e-5, | |
| "use_cache": true, | |
| "bos_token_id": 31999, | |
| "eos_token_id": 31999, | |
| "tokenizer_class": "GPTNeoXJapaneseTokenizer", | |
| "transformers_version": "4.19.0.dev0" | |
| } |