Demo script running issue

[liang315]

Do I need to modify the configuration file when running the demo script? I encountered this error

python demo.py --config-file .\configs\VLDet_OVCOCO_CLIP_R50_1x_caption.yaml --input .\heatmap_inputs\000000059598.jpg --output .\demo_outputs\ --vocabulary coco --opts MODEL.WEIGHTS models/coco_vldet.pth

Traceback (most recent call last):
  File "E:\VLDet\VLDet\demo.py", line 152, in <module>
    predictions, visualized_output = demo.run_on_image(img)
  File "E:\VLDet\VLDet\vldet\predictor.py", line 84, in run_on_image
    predictions = self.predictor(image)
  File "E:\VLDet\VLDet\detectron2\detectron2\engine\defaults.py", line 317, in __call__
    predictions = self.model([inputs])[0]
  File "E:\anaconda3\envs\VLDet\lib\site-packages\torch\nn\modules\module.py", line 1130, in _call_impl
    return forward_call(*input, **kwargs)
  File "E:\VLDet\VLDet\vldet\modeling\meta_arch\custom_rcnn.py", line 111, in forward
    return self.inference(batched_inputs)
  File "E:\VLDet\VLDet\vldet\modeling\meta_arch\custom_rcnn.py", line 96, in inference
    results, _ = self.roi_heads(images, features, proposals)
  File "E:\anaconda3\envs\VLDet\lib\site-packages\torch\nn\modules\module.py", line 1130, in _call_impl
    return forward_call(*input, **kwargs)
  File "E:\VLDet\VLDet\vldet\modeling\roi_heads\vldet_roi_heads.py", line 230, in forward
    features, proposals, classifier_info=classifier_info)
  File "E:\VLDet\VLDet\vldet\modeling\roi_heads\vldet_roi_heads.py", line 121, in _forward_box
    ann_type=ann_type, classifier_info=classifier_info)
  File "E:\VLDet\VLDet\vldet\modeling\roi_heads\vldet_roi_heads.py", line 331, in _run_stage
    classifier_info=classifier_info)
  File "E:\anaconda3\envs\VLDet\lib\site-packages\torch\nn\modules\module.py", line 1130, in _call_impl
    return forward_call(*input, **kwargs)
  File "E:\VLDet\VLDet\vldet\modeling\roi_heads\vldet_fast_rcnn.py", line 1604, in forward
    cls_scores, proj_region, zs_weight = self.cls_score(x, ann_type)
  File "E:\anaconda3\envs\VLDet\lib\site-packages\torch\nn\modules\module.py", line 1130, in _call_impl
    return forward_call(*input, **kwargs)
  File "E:\VLDet\VLDet\vldet\modeling\roi_heads\zero_shot_classifier.py", line 135, in forward
    x = torch.mm(proj_x, zs_weight)
RuntimeError: mat1 and mat2 shapes cannot be multiplied (256x1024 and 512x1204)