Migrating to torch 1.x

[ramilmsh]

I have updated the models to torch native (they are available through link in README.md) and modified the simplify.py accordingly. All models seem to work

[liuqk3]

@ramilmsh
Hi, thanks for your work.
Have you validated the converted models? And can you public your models for PyTorch 1.0+?
Thanks a lot

[ramilmsh]

Have you validated the converted models?

I have run several images on it and it seemed to work on all models.

And can you public your models for PyTorch 1.0+?

I have uploaded them on Mega, since I have no CDN account.

[Trass3r]

Note that the original models aren't even available anymore, at least atm.

[ramilmsh]

Note that the original models aren't even available anymore, at least atm.

I think I have a copy on my local machine, I will try to upload them to the same mega folder in a few days

[imcomking]

It works well at pytorch 1.5.1
thank you

[Xu-Justin]

I have tried this code. It's working!

[junguler]

just tried this and it works, thanks

[junguler]

just a quick question tho, how can i change the model? i've tried --model model_gan.pth but it throws out an error

Traceback (most recent call last):
  File "C:\git\sketch_simplification\simplify.py", line 15, in <module>
    model_import = __import__(opt.model, fromlist=['model', 'immean', 'imstd'])
ModuleNotFoundError: No module named 'model_gan.pth'; 'model_gan' is not a package

same error with the other 3 models

[Xu-Justin]

just a quick question tho, how can i change the model? i've tried --model model_gan.pth but it throws out an error

Traceback (most recent call last):
  File "C:\git\sketch_simplification\simplify.py", line 15, in <module>
    model_import = __import__(opt.model, fromlist=['model', 'immean', 'imstd'])
ModuleNotFoundError: No module named 'model_gan.pth'; 'model_gan' is not a package

same error with the other 3 models

1. Have you downloaded the model?
2. How do you run the code? (i.e. the command you use to run the code)

I have tried changing the model and it's working, could you provide more details to recreate the error?

[junguler]

I have tried changing the model and it's working, could you provide more details to recreate the error?

yes, i've downloaded the 4 .pth models from that mega link, i also have the 4 .t7 models that the download script downloads, both sets of files are at the root of the folder

i'm running this with miniconda on windows 10, here is the command i've tried to run

python simplify.py --img face.jpg --out test.png --model model_gan.pth

the program works fine without specifying model but i can't get it working with any of the 8 model files

[Xu-Justin]

yes, i've downloaded the 4 .pth models from that mega link, i also have the 4 .t7 models that the download script downloads, both sets of files are at the root of the folder

i'm running this with miniconda on windows 10, here is the command i've tried to run

python simplify.py --img face.jpg --out test.png --model model_gan.pth

the program works fine without specifying model but i can't get it working with any of the 8 model files

You shouldn't include .pth in the model arguments.

python simplify.py --img face.jpg --out test.png --model model_gan

[junguler]

You shouldn't include .pth in the model arguments.

worked perfectly, thank you for taking the time to explain this to me

[CrazedPrototype]

Hi,
I am encountering an issue when running this code using GAN model. The script runs without errors, but the output image is entirely white.

python: 3.7.0
pytorch: 1.5.1
cudatoolkit: 9.2.148
GPU: NVIDIA GeForce RTX 3050 ti
OS: Windows 11

[Xu-Justin]

Hi, I am encountering an issue when running this code using GAN model. The script runs without errors, but the output image is entirely white.

python: 3.7.0 pytorch: 1.5.1 cudatoolkit: 9.2.148 GPU: NVIDIA GeForce RTX 3050 ti OS: Windows 11

Can you provide the arguments you use to run the program?

[CrazedPrototype]

Hi, I am encountering an issue when running this code using GAN model. The script runs without errors, but the output image is entirely white.
python: 3.7.0 pytorch: 1.5.1 cudatoolkit: 9.2.148 GPU: NVIDIA GeForce RTX 3050 ti OS: Windows 11

Can you provide the arguments you use to run the program?

python simplify.py --img test.png --out test_out.png --model model_gan

All these are saved in the same folder as simplify.py. (within - E:\sketch\sketch_simplification-master)

Its also taking around 10 mins to give an output.

[YY-RR-ZZ]

Thank you very much for providing the code and the PTH weights.

[horatioh13]

works well for me.
Ubuntu 24.04, no GPU.
using Python 3.7.17
torch 1.13.1