RedHatAI
/

Qwen2.5-VL-72B-Instruct-FP8-Dynamic

Image-Text-to-Text

text-generation-inference

compressed-tensors

Model card Files Files and versions Community

zifeitong commited on 15 days ago

Commit

33512fd

·

verified ·

1 Parent(s): 0f467bb

Fix processor_class to match upstream

Otherwise vllm won't work correctly.

Files changed (1) hide show

preprocessor_config.json +1 -1

preprocessor_config.json CHANGED Viewed

@@ -18,7 +18,7 @@
   "min_pixels": 3136,
   "patch_size": 14,
   "image_processor_type": "Qwen2VLImageProcessor",
-  "processor_class": "Qwen2VLImageProcessor",
   "resample": 3,
   "rescale_factor": 0.00392156862745098,
   "size": {

   "min_pixels": 3136,
   "patch_size": 14,
   "image_processor_type": "Qwen2VLImageProcessor",
+  "processor_class": "Qwen2_5_VLProcessor",
   "resample": 3,
   "rescale_factor": 0.00392156862745098,
   "size": {