mehdidc/feed_forward_vqgan_clip
Feed forward VQGAN-CLIP model, where the goal is to eliminate the need for optimizing the latent space of VQGAN for each input prompt
PythonMIT
Issues
- 0
How to get more variation in the null image
#27 opened by kchodorow - 25
Positional Stickiness
#8 opened by afiaka87 - 2
How to condition model output z that looks like it can from a standard normal distribution?
#20 opened by xiankgx - 1
Repo License
#24 opened by minimaxir - 12
Models are broken in the new `torch` version
#25 opened by neverix - 3
Slow Training Speed
#21 opened by s13kman - 1
training GPU configuration
#23 opened by CrossLee1 - 9
Error in Load Model
#19 opened by metaphorz - 8
New Checkpoint Idea
#22 opened by afiaka87 - 1
- 2
clarifying differences between available models
#18 opened by zeke - 7
Unavailable and broken links
#15 opened by woctezuma - 2
- 5
CLIP-guided-diffusion updates
#9 opened by afiaka87 - 3
VQGAN - blended models
#12 opened by johndpope - 7
Not an issue - richer datasets
#6 opened by johndpope - 6
- 10
- 1
New CLIP checkpoints from Open AI
#4 opened by afiaka87