VictorMylle/Kohya-ss-sd-scripts
1
0
Fork 1
mirror of https://github.com/kohya-ss/sd-scripts.git synced 2026-04-09 06:45:09 +00:00
Code Issues Packages Projects Releases Wiki Activity

add sample image generation during training

Browse Source
This commit is contained in:
Kohya S
2024-08-14 22:15:26 +09:00
parent 56d7651f08
commit 7db4222119
4 changed files with 374 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
README.md
View File

@@ -6,6 +6,8 @@ This feature is experimental. The options and the training script may change in
__Please update PyTorch to 2.4.0. We have tested with PyTorch 2.4.0 with CUDA 12.4. We also updated `accelerate` to 0.33.0 just to be safe. `requirements.txt` is also updated, so please update the requirements.__
Aug 14, 2024: Sample image generation during training is now supported. Specify options such as `--sample_prompts` and `--sample_every_n_epochs`. It will be very slow when `--split_mode` is specified.
Aug 13, 2024:
__Experimental__ A network argument `train_blocks` is added to `lora_flux`. This is to select the target blocks of LoRA from FLUX double blocks and single blocks. Specify like `--network_args "train_blocks=single"`. `all` trains both double blocks and single blocks, `double` trains only double blocks, and `single` trains only single blocks. The default (omission) is `all`.