Age | Commit message (Collapse) | Author |
|
and kb
|
|
|
|
fix bug when using --gfpgan-models-path
|
|
Update prompts_from_file script to allow concatenating entries with the general prompt.
|
|
call state.jobnext() before postproces*()
|
|
Fix #13796
|
|
Fix parenthesis auto selection
|
|
|
|
requirements_versions httpx==0.24.1
|
|
requirements_versions httpx==0.24.1
|
|
|
|
Fixes #13813
|
|
|
|
diag_oft has MultiheadAttn which kohya's doesn't?, attempt create new module based off network_lora.py, errors about tensor dim mismatch
|
|
|
|
|
|
Fix comment error that makes understanding scheduling more confusing.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Since norm layer need fp32, I only convert the linear operation layer(conv2d/linear)
And TE have some pytorch function not support bf16 amp in CPU. I add a condition to indicate if the autocast is for unet.
|
|
|
|
general prompt.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
old-style emphasis
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|