Instructions to use google/gemma-4-12B-it-qat-w4a16-ct with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use google/gemma-4-12B-it-qat-w4a16-ct with Transformers:
# Load model directly from transformers import AutoProcessor, AutoModelForMultimodalLM processor = AutoProcessor.from_pretrained("google/gemma-4-12B-it-qat-w4a16-ct") model = AutoModelForMultimodalLM.from_pretrained("google/gemma-4-12B-it-qat-w4a16-ct") - Notebooks
- Google Colab
- Kaggle
fix: add missing num_soft_tokens and model_patch_size to vision_config
#5
by lucianommartins - opened
- config.json +3 -1
config.json
CHANGED
|
@@ -214,7 +214,9 @@
|
|
| 214 |
},
|
| 215 |
"mm_embed_dim": 3840,
|
| 216 |
"mm_posemb_size": 1120,
|
|
|
|
| 217 |
"model_type": "gemma4_unified_vision",
|
|
|
|
| 218 |
"output_attentions": false,
|
| 219 |
"output_hidden_states": false,
|
| 220 |
"output_proj_dims": 3840,
|
|
@@ -224,4 +226,4 @@
|
|
| 224 |
"return_dict": true,
|
| 225 |
"rms_norm_eps": 1e-06
|
| 226 |
}
|
| 227 |
-
}
|
|
|
|
| 214 |
},
|
| 215 |
"mm_embed_dim": 3840,
|
| 216 |
"mm_posemb_size": 1120,
|
| 217 |
+
"model_patch_size": 48,
|
| 218 |
"model_type": "gemma4_unified_vision",
|
| 219 |
+
"num_soft_tokens": 280,
|
| 220 |
"output_attentions": false,
|
| 221 |
"output_hidden_states": false,
|
| 222 |
"output_proj_dims": 3840,
|
|
|
|
| 226 |
"return_dict": true,
|
| 227 |
"rms_norm_eps": 1e-06
|
| 228 |
}
|
| 229 |
+
}
|