Mirrors
/
EveryDream2trainer
mirror of https://github.com/victorchall/EveryDream2trainer.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Update ADVANCED_TWEAKING.md

This commit is contained in:
Victor Hall 2023-07-16 17:57:17 -04:00 committed by GitHub
parent a6032bee35
commit 28e3d22aff
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 4 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
doc/ADVANCED_TWEAKING.md
View File

@ -151,6 +151,10 @@ Test results: https://huggingface.co/panopstor/ff7r-stable-diffusion/blob/main/z
Very tentatively, I suggest closer to 0.10 for short term training, and lower values of around 0.02 to 0.03 for longer runs (50k+ steps). Early indications seem to suggest values like 0.10 can cause divergance over time.
## Zero terminal SNR
Set `zero_frequency_noise_ratio` to -1.
## Keeping images together (custom batching)
If you have a subset of your dataset that expresses the same style or concept, training quality may be improved by putting all of these images through the trainer together in the same batch or batches, instead of the default behaviour (which is to shuffle them randomly throughout the entire dataset).