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

update README and clean up code for schedulefree optimizer

Browse Source
This commit is contained in:
Kohya S
2024-12-01 22:00:44 +09:00
parent 14c9ba925f
commit 1dc873d9b4
2 changed files with 6 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
README.md
View File

@@ -16,9 +16,11 @@ The command to install PyTorch is as follows:
1 Dec, 2024:
- Pseudo Huber loss is now available for FLUX.1 and SD3.5 training. See [#1808](https://github.com/kohya-ss/sd-scripts/pull/1808) for details. Thanks to recris!
- Pseudo Huber loss is now available for FLUX.1 and SD3.5 training. See PR [#1808](https://github.com/kohya-ss/sd-scripts/pull/1808) for details. Thanks to recris!
- Specify `--loss_type huber` or `--loss_type smooth_l1` to use it. `--huber_c` and `--huber_scale` are also available.
- [Prodigy + ScheduleFree](https://github.com/LoganBooker/prodigy-plus-schedule-free) is supported. See PR [#1811](https://github.com/kohya-ss/sd-scripts/pull/1811) for details. Thanks to rockerBOO!
Nov 14, 2024:
- Improved the implementation of block swap and made it available for both FLUX.1 and SD3 LoRA training. See [FLUX.1 LoRA training](#flux1-lora-training) etc. for how to use the new options. Training is possible with about 8-10GB of VRAM.

5
library/train_util.py
View File

@@ -4883,7 +4883,6 @@ def get_optimizer(args, trainable_params):
optimizer = optimizer_class(trainable_params, lr=lr, **optimizer_kwargs)
elif optimizer_type.endswith("schedulefree".lower()):
should_train_optimizer = True
try:
import schedulefree as sf
except ImportError:
@@ -5000,8 +4999,8 @@ def get_optimizer(args, trainable_params):
optimizer_name = optimizer_class.__module__ + "." + optimizer_class.__name__
optimizer_args = ",".join([f"{k}={v}" for k, v in optimizer_kwargs.items()])
if hasattr(optimizer, 'train') and callable(optimizer.train):
# make optimizer as train mode: we don't need to call train again, because eval will not be called in training loop
if hasattr(optimizer, "train") and callable(optimizer.train):
# make optimizer as train mode before training for schedulefree optimizer. the optimizer will be in eval mode in sampling and saving.
optimizer.train()
return optimizer_name, optimizer_args, optimizer