mirror of
https://github.com/kohya-ss/sd-scripts.git
synced 2026-04-09 06:45:09 +00:00
Move freeze_blocks to sd3_train because it's only for sd3
This commit is contained in:
@@ -309,6 +309,9 @@ resolution = [512, 512]
|
||||
|
||||
SD3 training is done with `sd3_train.py`.
|
||||
|
||||
__Sep 1, 2024__:
|
||||
- `--num_last_block_to_freeze` is added to `sd3_train.py`. This option is to freeze the last n blocks of the MMDiT. See [#1417](https://github.com/kohya-ss/sd-scripts/pull/1417) for details. Thanks to sdbds!
|
||||
|
||||
__Jul 27, 2024__:
|
||||
- Latents and text encoder outputs caching mechanism is refactored significantly.
|
||||
- Existing cache files for SD3 need to be recreated. Please delete the previous cache files.
|
||||
|
||||
Reference in New Issue
Block a user