CoolFace
Modelpublic

Scale-or-Reason/gemma3-4B_0_split

sourceHugging Faceupdated 10mo agoView on Hugging Face
0likes5downloads
config.json94 linesDownload Raw Back to root
1{2  "architectures": [3    "Gemma3ForConditionalGeneration"4  ],5  "boi_token_index": 255999,6  "eoi_token_index": 256000,7  "image_token_index": 262144,8  "initializer_range": 0.02,9  "mm_tokens_per_image": 256,10  "model_type": "gemma3",11  "text_config": {12    "_sliding_window_pattern": 6,13    "attention_bias": false,14    "attention_dropout": 0.0,15    "attn_logit_softcapping": null,16    "final_logit_softcapping": null,17    "head_dim": 256,18    "hidden_activation": "gelu_pytorch_tanh",19    "hidden_size": 2560,20    "initializer_range": 0.02,21    "intermediate_size": 10240,22    "layer_types": [23      "sliding_attention",24      "sliding_attention",25      "sliding_attention",26      "sliding_attention",27      "sliding_attention",28      "full_attention",29      "sliding_attention",30      "sliding_attention",31      "sliding_attention",32      "sliding_attention",33      "sliding_attention",34      "full_attention",35      "sliding_attention",36      "sliding_attention",37      "sliding_attention",38      "sliding_attention",39      "sliding_attention",40      "full_attention",41      "sliding_attention",42      "sliding_attention",43      "sliding_attention",44      "sliding_attention",45      "sliding_attention",46      "full_attention",47      "sliding_attention",48      "sliding_attention",49      "sliding_attention",50      "sliding_attention",51      "sliding_attention",52      "full_attention",53      "sliding_attention",54      "sliding_attention",55      "sliding_attention",56      "sliding_attention"57    ],58    "max_position_embeddings": 131072,59    "model_type": "gemma3_text",60    "num_attention_heads": 8,61    "num_hidden_layers": 34,62    "num_key_value_heads": 4,63    "query_pre_attn_scalar": 256,64    "rms_norm_eps": 1e-06,65    "rope_local_base_freq": 10000.0,66    "rope_scaling": {67      "factor": 8.0,68      "rope_type": "linear"69    },70    "rope_theta": 1000000.0,71    "sliding_window": 1024,72    "torch_dtype": "bfloat16",73    "use_cache": false,74    "vocab_size": 26220875  },76  "torch_dtype": "bfloat16",77  "transformers_version": "4.55.2",78  "vision_config": {79    "attention_dropout": 0.0,80    "hidden_act": "gelu_pytorch_tanh",81    "hidden_size": 1152,82    "image_size": 896,83    "intermediate_size": 4304,84    "layer_norm_eps": 1e-06,85    "model_type": "siglip_vision_model",86    "num_attention_heads": 16,87    "num_channels": 3,88    "num_hidden_layers": 27,89    "patch_size": 14,90    "torch_dtype": "bfloat16",91    "vision_use_head": false92  }93}94