Verified: 2 years ago
SafeTensor
Type | |
Stats | 1,080 8,654 |
Reviews | (138) |
Published | Mar 30, 2023 |
Base Model | |
Training | Epochs: 10 |
Hash | AutoV2 54223CA615 |
At attempt at creating flash photography lighting.
Trained on 362 images: (People on street, at parties, animal/nature photography, video game screenshots, and some traditional art.)
But it isn't consistent or strong enough. And it might have baked in some of the human models faces?
Not sure how to improve it. I think I will try to find more photos of just people and remove the non human centric photos.
If you know where I can find some, please share.