Skip to content

Fix issue where LoRa loads on a device different from Dit #1171

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
Artiprocher merged 1 commit into from
Jan 5, 2026
Merged

Fix issue where LoRa loads on a device different from Dit #1171

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
2 changes: 2 additions & 0 deletions diffsynth/utils/lora/flux.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -149,6 +149,8 @@ def guess_alpha(state_dict):
dtype=state_dict_[name].dtype)
else:
state_dict_.pop(name.replace(".a_to_q.", ".proj_in_besides_attn."))

mlp = mlp.to(device=state_dict_[name].device)
Copy link
Contributor

@gemini-code-assist gemini-code-assist bot Dec 31, 2025

Choose a reason for hiding this comment

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

medium

This change introduces a small inefficiency. When mlp is created with torch.zeros (in the if mlp is None branch), it's created on the CPU and then moved to the target device here. This causes an unnecessary data copy. For better performance, it would be ideal to create the tensor directly on the target device by adding device=state_dict_[name].device to the torch.zeros call on line 147. This .to() call would then only be needed for the else branch.

if 'lora_A' in name:
param = torch.concat([
state_dict_.pop(name),
Expand Down