KeyError: 'patch_embed.proj.weight' about video-swin-transformer

[r-kellerm]

Hi, please see this issue #21, you need to install our project instead of mmaction2 repo.

Thanks for your response!
uninstalled mmaction2 and reinstalled with pip install -v -e . , but still getting the error from mmcv:
2021-09-30 13:20:02,304 - mmaction - INFO - load model from: pretrained/swin_small_patch244_window877_kinetics400_1k.pth
Traceback (most recent call last):

File "/.local/lib/python3.8/site-packages/mmcv/utils/registry.py", line 52, in build_from_cfg
return obj_cls(**args)
File "exp/Video-Swin-Transformer/mmaction/models/recognizers/base.py", line 109, in init
self.init_weights()
File "exp/Video-Swin-Transformer/mmaction/models/recognizers/base.py", line 126, in init_weights
self.backbone.init_weights()
File "exp/Video-Swin-Transformer/mmaction/models/backbones/swin_transformer.py", line 641, in init_weights
self.inflate_weights(logger)
File "exp/Video-Swin-Transformer/mmaction/models/backbones/swin_transformer.py", line 588, in inflate_weights
state_dict['patch_embed.proj.weight'] = state_dict['patch_embed.proj.weight'].unsqueeze(2).repeat(1,1,self.patch_size[0],1,1) / self.patch_size[0]
KeyError: 'patch_embed.proj.weight'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "tools/train.py", line 201, in
main()
File "tools/train.py", line 156, in main
model = build_model(
File "exp/Video-Swin-Transformer/mmaction/models/builder.py", line 70, in build_model
return build_localizer(cfg)
File "exp/Video-Swin-Transformer/mmaction/models/builder.py", line 62, in build_localizer
return LOCALIZERS.build(cfg)
File "/.local/lib/python3.8/site-packages/mmcv/utils/registry.py", line 212, in build
return self.build_func(*args, **kwargs, registry=self)
File "/.local/lib/python3.8/site-packages/mmcv/cnn/builder.py", line 27, in build_model_from_cfg
return build_from_cfg(cfg, registry, default_args)
File "/.local/lib/python3.8/site-packages/mmcv/utils/registry.py", line 55, in build_from_cfg
raise type(e)(f'{obj_cls.name}: {e}')
KeyError: "Recognizer3D: 'patch_embed.proj.weight'"

Installed mmcv through: pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/{cu_version}/{torch_version}/index.html.

[Dzw1219]

Modified the config file.
Add 'pretrained2d=False'
model=dict(backbone=dict(patch_size=(2,4,4), drop_path_rate=0.1, pretrained2d=False), test_cfg=dict(max_testing_views=4))

[hust-nj]

Hi, please see this issue #21, you need to install our project instead of mmaction2 repo.