[XPU][LoRA] Support LoRA for DeepSeek V4 on XPU - #53689
Merged
Merged
Conversation
Port the NVIDIA DSV4 LoRA support (vllm-project#53361) to the XPU model: add SupportsLoRA, packed_modules_mapping, and lora_skip_prefixes to DeepseekV4ForCausalLM; reconcile LoRA-wrapped (.base_layer.) parameter names in load_weights; and add the matching .base_layer. scale regex variants to the weights mapper. Signed-off-by: Chaojun Zhang <chaojun.zhang@intel.com>
![claude[bot]](https://avatars.githubusercontent.com/in/1236702?s=60&v=4){kind=small}
mergify
Bot
added
deepseek
Contributor
|
@chaojun-zhang can you provide test command for this support? |
jeejeelee
approved these changes
Aug 25, 2026
Contributor
Author
The original #53361 PR was validated through an external verl end-to-end test, not via an in-repo script. Also, there is no publicly available LoRA adapter for testing at the moment. We can set this aside for now. |
Member
Contributor
Author
|
/ci run |
|
✅ Triggered Buildkite CI #87045 for commit |
Contributor
Author
|
/ci retry |
|
✅ Queued 6 failed job(s) for retry in Buildkite CI #87045. |
ItsRoy69
pushed a commit
to ItsRoy69/vllm
that referenced
this pull request
Sep 10, 2026
Signed-off-by: Chaojun Zhang <chaojun.zhang@intel.com> Signed-off-by: Jyotirmoy Roy <jyotirmoyroy649@gmail.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Follow-up to #53361, which added LoRA support for DeepSeek V4 but only touched
vllm/models/deepseek_v4/nvidia/model.py. This ports the same changes tovllm/models/deepseek_v4/xpu/model.pyTest plan: #53361 added no repo tests (validated via external verl E2E), and there's no existing DeepSeek V4 + LoRA coverage in
tests/lora/ortests/models/deepseek_v4/to extend here either. If upstream later adds repo tests for DSV4+LoRA, they should be extended to run on XPU too.