Frequently Asked Questions (常见问题）

[Zeqiang-Lai]

Q: Expecting value: line 1 column 1 (char 0)

A: This might be caused by network proxy. Please refer to #50 for solutions.

• Expecting value: line 1 column 1 (char 0) #50

[Zeqiang-Lai]

Q: RuntimeError: PytorchStreamReader fail reading zip archive

A: The checkpoint is broken. Download again by removing the broken one. Optionally, you could download the checkpoint manually from here, and put it to the target directory.

The target directory could be found via

python -c "import draggan; print(draggan.home)"

• RuntimeError: PytorchStreamReader failed reading zip archive: failed finding central directory #48

[Zeqiang-Lai]

StyleGAN Custom OP, DLL not found

• DLL load failed #12
• Failed to load Torch extension #40
• DLL load failed while importing fused: The specified module could not be found. #38
• Fused extension build error #31
• ninja: build stopped: subcommand failed.,how to fix it| Ubuntu 22.04 #35
• Error: warnings.warn(f'Error checking compiler version for {compiler}: {error}') #30

[Zeqiang-Lai]

Q: No Cuda Runtime

• Error launching after installing requirements #32
• AssertionError: Torch not compiled with CUDA enabled #51
• Cuda #37

[Zeqiang-Lai]

Q: Requirements

• Could you give out the version of requirements? #11
• Which torch is supported?  #10
• Requirements Error #13

[Zeqiang-Lai]

Q: Does this project support uploading custom image ?

A: Yes, this project support custom image.

[Zeqiang-Lai]

Performance

• very slow #52
• Out of memory error #44

[Zeqiang-Lai]

Q: The custom image is of low quality, distorted, fail to drag.

A. The results might be distorted, this is an issue of the algorithms instead of the implementation. More advanced algorithms will be integrated in the future.

• 'NoneType' object has no attribute 'rsplit'  #54
• Woman face collapsed after rotation #26
• GAN conversion distorted #25
• Lion face changed but distorted #24
• Lion face fully distorted at the load of image #23
• Lion head didn't move and mouth didn't open. #22
• The nose has been larger and the face has distorted. #19
• Uploading a new image will result in a collapsed image. What should I do? #53

[Zeqiang-Lai]

Hardware Requirements

[Zeqiang-Lai]

Q: AttributeError: module torchvision.transforms has no attribute "Resize"

A: Then version of torchvision is too low. Upgrade it via

pip install torchvision --upgrade