Skip to content

Sreerag-ibtl/inference_AttnGan_py3

Repository files navigation

inference_AttnGan_py3

AttnGAN is a work I found here. https://github.com/taoxugit/AttnGAN

It can convert text descriptions into quality images when trained on a dataset.

Only the GUI section is coded by me. All other parts were converted from original codebase in Python2 to Python3 using 2to3. All the models were obtained from the official repo. Some parts has been edited to get compatible with PyTorch cpu only.

I used anaconda for development in a windows 10 machine. Python version was 3.8.2. PyTorch CPU is used instead of PyTorch GPU.The conda package list provided in the requirements.txt.

Result from the GUI:

Save the text encoder into DAMSMencoders from https://drive.google.com/open?id=1GNUKjVeyWYBJ8hEU-yrfYQpDOkxEyP3V

Save the model into models from https://drive.google.com/open?id=1lqNG75suOuR_8gjoEPYNp8VyT_ufPPig

Install packages specified in requirements.txt

Move into the directory code.

Run

python gui.py

About

A GUI for getting prediction from https://github.com/taoxugit/AttnGAN with single caption input.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages