openfree/Gemma-3-R1984-4B

1 2{3  "architectures": [4    "Gemma3ForConditionalGeneration"5  ],6  "boi_token_index": 255999,7  "eoi_token_index": 256000,8  "eos_token_id": [9    1,10    10611  ],12  "image_token_index": 262144,13  "initializer_range": 0.02,14  "mm_tokens_per_image": 256,15  "model_type": "gemma3",16  "text_config": {17    "attention_bias": false,18    "attention_dropout": 0.0,19    "attn_logit_softcapping": null,20    "cache_implementation": "hybrid",21    "final_logit_softcapping": null,22    "head_dim": 256,23    "hidden_activation": "gelu_pytorch_tanh",24    "hidden_size": 2560,25    "initializer_range": 0.02,26    "intermediate_size": 10240,27    "max_position_embeddings": 131072,28    "model_type": "gemma3_text",29    "num_attention_heads": 8,30    "num_hidden_layers": 34,31    "num_key_value_heads": 4,32    "query_pre_attn_scalar": 256,33    "rms_norm_eps": 1e-06,34    "rope_local_base_freq": 10000.0,35    "rope_scaling": {36      "factor": 8.0,37      "rope_type": "linear"38    },39    "rope_theta": 1000000.0,40    "sliding_window": 1024,41    "sliding_window_pattern": 6,42    "torch_dtype": "bfloat16",43    "use_cache": true,44    "vocab_size": 26220845  },46  "torch_dtype": "bfloat16",47  "transformers_version": "4.50.0.dev0",48  "vision_config": {49    "attention_dropout": 0.0,50    "hidden_act": "gelu_pytorch_tanh",51    "hidden_size": 1152,52    "image_size": 896,53    "intermediate_size": 4304,54    "layer_norm_eps": 1e-06,55    "model_type": "siglip_vision_model",56    "num_attention_heads": 16,57    "num_channels": 3,58    "num_hidden_layers": 27,59    "patch_size": 14,60    "torch_dtype": "bfloat16",61    "vision_use_head": false62  }63}64