Skip to content

Fix gradient-checkpointing option is ignored in SDXL+LoRA training. (#6388) #6402

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
sayakpaul merged 5 commits into from
Jan 1, 2024
Merged

Fix gradient-checkpointing option is ignored in SDXL+LoRA training. (#6388) #6402

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
091c589
Fix gradient-checkpointing option is ignored in SDXL+LoRA training. (…
2510 Dec 30, 2023
07a5219
Fix gradient-checkpointing option is ignored in SD+LoRA training.
2510 Dec 30, 2023
0e9b7b1
Merge branch 'main' into fix-gradient-checkpointing-ignored-in-sdxl-lora
sayakpaul Dec 31, 2023
8d7feb6
Fix gradient checkpoint is not applied to text encoders. (SDXL+LoRA)
2510 Dec 31, 2023
fc8f1fe
Merge branch 'main' into fix-gradient-checkpointing-ignored-in-sdxl-lora
sayakpaul Dec 31, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions examples/text_to_image/train_text_to_image_lora.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -486,6 +486,9 @@ def main():

lora_layers = filter(lambda p: p.requires_grad, unet.parameters())

if args.gradient_checkpointing:
unet.enable_gradient_checkpointing()

# Enable TF32 for faster training on Ampere GPUs,
# cf https://pytorch.org/docs/stable/notes/cuda.html#tensorfloat-32-tf32-on-ampere-devices
if args.allow_tf32:
Expand Down
6 changes: 6 additions & 0 deletions examples/text_to_image/train_text_to_image_lora_sdxl.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -706,6 +706,12 @@ def load_model_hook(models, input_dir):
accelerator.register_save_state_pre_hook(save_model_hook)
accelerator.register_load_state_pre_hook(load_model_hook)

if args.gradient_checkpointing:
unet.enable_gradient_checkpointing()
Comment on lines +709 to +710
Copy link
Copy Markdown
Member

@sayakpaul sayakpaul Dec 31, 2023

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The text encoder can be trainable in this script too:

diffusers/examples/text_to_image/train_text_to_image_lora_sdxl.py

Line 621 in bf725e0

if args.train_text_encoder:

This is how it's done:

diffusers/examples/dreambooth/train_dreambooth_lora_sdxl.py

Line 976 in bf725e0

text_encoder_one.gradient_checkpointing_enable()

if args.train_text_encoder:
text_encoder_one.gradient_checkpointing_enable()
text_encoder_two.gradient_checkpointing_enable()

# Enable TF32 for faster training on Ampere GPUs,
# cf https://pytorch.org/docs/stable/notes/cuda.html#tensorfloat-32-tf32-on-ampere-devices
if args.allow_tf32:
Expand Down