Pytorch and torchvision version depreciated

[chaudharykapil]

conda install pytorch=0.4.1 torchvision=0.2.1 -c pytorch not working

Result:

Collecting package metadata (current_repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Collecting package metadata (repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.

PackagesNotFoundError: The following packages are not available from current channels:

• pytorch=0.4.1
• torchvision=0.2.1

Current channels:

• https://conda.anaconda.org/pytorch/win-64
• https://conda.anaconda.org/pytorch/noarch
• https://repo.anaconda.com/pkgs/main/win-64
• https://repo.anaconda.com/pkgs/main/noarch
• https://repo.anaconda.com/pkgs/r/win-64
• https://repo.anaconda.com/pkgs/r/noarch
• https://repo.anaconda.com/pkgs/msys2/win-64
• https://repo.anaconda.com/pkgs/msys2/noarch

To search for alternate channels that may provide the conda package you're
looking for, navigate to

https://anaconda.org

and use the search bar at the top of the page.

[ARajgor]

I upgrade the project to a new version of libraries.

code is tested on the new version of torch=1.13.1, torchvision=0.14.1 using python 3.9 pip installation.

[March 2023 update: code is compatible with torch 2.0.0+cu117, torchaudio 2.0.1+cu117, torchvision 0.15.1+cu117 on python 3.9.13]

repo: https://github.com/ARajgor/cp-vton-plus

[cskyy]

Hello, do you know how to make test_mask?

[ARajgor]

you can refer my other repo for this, https://github.com/ARajgor/VTryon-flask-new it is based on Flask App for custom images.

[cskyy]

you can refer my other repo for this, https://github.com/ARajgor/VTryon-flask-new it is based on Flask App for custom images.

Hello, can you help me to see why the code after I changed did not generate results saved in the specified path, after running no error?

[ARajgor]

for the body mask, you need to parse images into png format. like the code mentioned on line 65 in application.py. that generate two files and saved them in image-parse and image-parse-new. after that, we pass files into the make_body_mask function with file_name.jpg and file_name.png.

the function takes the file_name.jpg from image folder and file_name.png from image-parse-new folder.

[cskyy]

for the body mask, you need to parse images into png format. like the code mentioned on line 65 in application.py. that generate two files and saved them in image-parse and image-parse-new. after that, we pass files into the make_body_mask function with file_name.jpg and file_name.png.

the function takes the file_name.jpg from image folder and file_name.png from image-parse-new folder.

Thank you very much for your help, I have successfully made the image_mask! Wish you good health and smooth work!

[cskyy]

Hello, I'm sorry to bother you again, but I have encountered a new problem. When I used relevant codes to process images of 1024*768, the following situation occurred, image-mask was blank. I would like to ask whether it is related to the resolution of the image? What needs to be changed?

[ARajgor]

Actually, I didn't change the original resolution that came with the original project but it is 192 X 256 px. the image parser inference.py also generates 192 X 256 px image and you have to pass it to the function. so may be there is an error of resolution.

[cskyy]

Actually, I didn't change the original resolution that came with the original project but it is 192 X 256 px. the image parser inference.py also generates 192 X 256 px image and you have to pass it to the function. so may be there is an error of resolution.

Ok, thank you very much for your help!