predict_for_dir.py如何调用

[jxncyym]

我的命令如下：
python3 scripts/predict_for_dir.py DucoNet ./DucoNet256.pth --images ./predict_images --masks ./predict_masks --gpu 7 --results-path ./DucoNet_generate

出现如下错误：
{'MODELS_PATH': './', 'EXPS_PATH': PosixPath('checkpoints/output'), 'HFLICKR_PATH': '/home/dataset/IHD/HFlickr', 'HDAY2NIGHT_PATH': '/home/dataset/IHD/Hday2night', 'HCOCO_PATH': '/home/dataset/IHD/HCOCO', 'HADOBE5K1_PATH': '/data/IHD/HAdobe5k', 'HADOBE5K_PATH': '/home/dataset/IHD/HAdobe5k_resized1024', 'IMAGENET_PRETRAINED_MODELS': {'HRNETV2_W18_SMALL': './pretrained_models/hrnet_w18_small_model_v2.pth', 'HRNETV2_W18': './pretrained_models/hrnetv2_w18_imagenet_pretrained.pth', 'HRNETV2_W32': './pretrained_models/hrnetv2_w32_imagenet_pretrained.pth', 'HRNETV2_W40': './pretrained_models/hrnetv2_w40_imagenet_pretrained.pth', 'HRNETV2_W48': './pretrained_models/hrnetv2_w48_imagenet_pretrained.pth', 'HistNet': './checkpoints/last_model/hen_train/last_checkpoint.pth'}, 'RESULTS_PATH': PosixPath('DucoNet_generate')}
/home/yuyiming/miniconda3/envs/py39/lib/python3.9/site-packages/torch/functional.py:507: UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at ../aten/src/ATen/native/TensorShape.cpp:3549.)
return _VF.meshgrid(tensors, **kwargs) # type: ignore[attr-defined]
Load checkpoint from path: ./DucoNet256.pth
Save images to DucoNet_generate
0%| | 0/4 [00:01<?, ?it/s]
Traceback (most recent call last):
File "/data2/yuyiming/harmonization/DucoNet-Image-Harmonization/scripts/predict_for_dir.py", line 107, in
main()
File "/data2/yuyiming/harmonization/DucoNet-Image-Harmonization/scripts/predict_for_dir.py", line 60, in main
pred = predictor.predict(image, mask, composite_image_lab)
File "/data2/yuyiming/harmonization/DucoNet-Image-Harmonization/./iharm/inference/predictor.py", line 39, in predict
predicted_output = self.net(image, mask,image_lab = comp_image_lab)
File "/home/yuyiming/miniconda3/envs/py39/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1511, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/home/yuyiming/miniconda3/envs/py39/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1520, in _call_impl
return forward_call(*args, **kwargs)
File "/data2/yuyiming/harmonization/DucoNet-Image-Harmonization/./iharm/model/base/DucoNet_model.py", line 74, in forward
w_l = self.l_encoder(image_lab[:,0,:,:].unsqueeze(1),mask)
File "/home/yuyiming/miniconda3/envs/py39/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1511, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/home/yuyiming/miniconda3/envs/py39/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1520, in _call_impl
return forward_call(*args, **kwargs)
File "/data2/yuyiming/harmonization/DucoNet-Image-Harmonization/./iharm/model/base/Control_encoder.py", line 40, in forward
x_lab = torch.cat((x_lab, mask), dim=1)
RuntimeError: Sizes of tensors must match except in dimension 1. Expected size 1080 but got size 256 for tensor number 1 in the list.