Verified: 2 years ago
SafeTensor
Type | |
Stats | 932 9,118 |
Reviews | (133) |
Published | Jul 13, 2023 |
Base Model | |
Training | Steps: 20,700 Epochs: 30 |
Usage Tips | Clip Skip: 1 |
Hash | AutoV2 05594214E8 |
A LoRA trained from 207 carefully edited photos from the Library of Congress war photo collection (https://www.loc.gov/pictures/collection/fsac/). Images range from portraits of people to the machines of the 1944 industry. I tried to be really careful to not oversharpen or blowout values in any of the shots so that the resulting renders would be as high dynamic range as possible but your mileage may vary by the base model you might be using.