Skip to content

Fix lora tests failure in TPU CI due to the removal of LoRA bias #26723

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
DarkLight1337 merged 2 commits into from
Oct 14, 2025
Merged

Fix lora tests failure in TPU CI due to the removal of LoRA bias #26723

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
f3a9865
Fix lora test
vanbasten23 Oct 13, 2025
8a0f2af
remove unused args
vanbasten23 Oct 13, 2025
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: 1 addition & 2 deletions vllm/v1/worker/tpu_model_runner.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2128,12 +2128,11 @@ def _tpu_set_lora(
lora_a: torch.Tensor,
lora_b: torch.Tensor,
embeddings_tensor: torch.Tensor | None,
bias: torch.Tensor | None = None,
):
# TODO: The integer index leads to a recompilation, but converting it
# to a tensor doesn't seem to work anymore. This might be fixed with a
# later release of torch_xla.
self._original_set_lora(index, lora_a, lora_b, embeddings_tensor, bias)
self._original_set_lora(index, lora_a, lora_b, embeddings_tensor)
torch_xla.sync(wait=False)

def _tpu_reset_lora(self, index: int):
Expand Down