Sakalti/SJT-2B

1{2  "_name_or_path": "google/gemma-2-2b-jpn-it",3  "architectures": [4    "Gemma2ForCausalLM"5  ],6  "attention_bias": false,7  "attention_dropout": 0.0,8  "attn_logit_softcapping": 50.0,9  "bos_token_id": 2,10  "cache_implementation": "hybrid",11  "dtype": "bfloat16",12  "eos_token_id": 1,13  "final_logit_softcapping": 30.0,14  "head_dim": 256,15  "hidden_activation": "gelu_pytorch_tanh",16  "hidden_size": 2304,17  "initializer_range": 0.02,18  "intermediate_size": 9216,19  "max_position_embeddings": 131072,20  "model_type": "gemma2",21  "num_attention_heads": 8,22  "num_hidden_layers": 26,23  "num_key_value_heads": 4,24  "pad_token_id": 0,25  "query_pre_attn_scalar": 224,26  "rms_norm_eps": 1e-06,27  "rope_scaling": {28    "factor": 16.0,29    "type": "linear"30  },31  "rope_theta": 10000.0,32  "sliding_window": 4096,33  "torch_dtype": "float16",34  "transformers_version": "4.47.1",35  "unsloth_version": "2025.1.5",36  "use_cache": true,37  "vocab_size": 25600038}39