Looking for contributors?
nbardy opened this issue ยท 5 comments
๐ Hey I was looking to scale up some MUSE runs as well. Just saw this is going on. Would be happy to help out.
@nbardy For sure! @patil-suraj might be a better person to coordinate with but there is a checklist at the end of the repo that has a bunch of tasks that he thought of! Also, we added some ideas we had to the issues so feel free to go ham.
Let us know if you have any questions!
Looks like the checklist is a little out of date seems @patil-suraj is already stated on T5.
This looks like a good place for me to start.
Train improved VQGANs models.
- Add training component models for VQGAN (EMA, discriminator, LPIPS etc).
- VGQAN training script.
Can look into setting up some eval sets as well.
- Collect different prompts for intermediate evaluations (Can reuse the prompts for dalle-mini, parti-prompts).
@nbardy Ah, sry. Am already working on a vqgan training script which is in the pr! Let me know if you think anything is missing. And intermediate eval prompts sound awesome!