search

gligen / GLIGEN

Open-Set Grounded Text-to-Image Generation
MIT License
2.02k stars 151 forks source link

Docker env not work, and need few modifications to run the code #87

Open iLori-Jiang opened 4 months ago

iLori-Jiang commented 4 months ago

First of all, thanks for your amazing job.

However, when I try to run the code, I failed to use the docker environment that you provided.

After a long time struggling, I took modifications suggested by (1) https://github.com/gligen/GLIGEN/pull/66 (2) https://github.com/gligen/GLIGEN/pull/33

Then build the conda environment from ./demo/environment.yaml, where I also met a package version conflict and I changed the NumPy to 'numpy==1.23.5' version.

What's more, the model that you provided in the HF are of '.bin' format, while in the 'gligen_inference.py' it will read the model in '.pth' format. What I did is to rename the model by the code, and change the reading format directly to '.bin' in the code.

After all these modifications, I can successfully run the inference and yield results. It would be nice if you could update the README to make the tutorial more clear, and I hope my struggle can help other people.

snehalmahajan2912 commented 4 weeks ago

How did you use the .bin file then? Change the .pth to .bin or something else. I am trying to train the model, but it's in .bin format and the code has .ckpt format in the code. Could you please help me?