Skip to content
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

Jump to bottom

[Bugfix] add input embedding #11684

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
Open

[Bugfix] add input embedding #11684

wants to merge 2 commits into from
+114 −4

Conversation

Bryce1010
Copy link
Contributor

@Bryce1010 Bryce1010 commented Jan 2, 2025 •
edited by github-actions bot
Loading

adds support for passing prompt_embeds to LLM.generate as

llm.generate({"prompt_embeds": input_embeds}, sampling_params)

or

llm.generate(
    [{"prompt_embeds": input_embeds} for input_embeds in inputs_embeds], sampling_params
)

this enables use cases when only the embedding layer is finetuned, and have the same model backend support multiple custom tuned embedding layers

FIX #416
FIX #8323

@github-actions GitHub Actions
Copy link

github-actions bot commented Jan 2, 2025

👋 Hi! Thank you for contributing to the vLLM project.
Just a reminder: PRs would not trigger full CI run by default. Instead, it would only run fastcheck CI which starts running only a small and essential subset of CI tests to quickly catch errors. You can run other CI tests on top of those by going to your fastcheck build on Buildkite UI (linked in the PR checks section) and unblock them. If you do not have permission to unblock, ping simon-mo or khluu to add you in our Buildkite org.

Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging.

To run CI, PR reviewers can do one of these:

  • Add ready label to the PR
  • Enable auto-merge.

🚀

@mergify mergify bot added the frontend label Jan 2, 2025
@Bryce1010 Bryce1010 force-pushed the feature/vllm/add-input-embedding branch 5 times, most recently from 0429875 to 8c7751d Compare January 2, 2025 11:17
@DarkLight1337
Copy link
Member

Thanks for opening this PR, can you explain what this PR is about and how this is related to #11375?

@Bryce1010
Copy link
Contributor Author

Bryce1010 commented Jan 2, 2025 •
edited
Loading

Thanks for opening this PR, can you explain what this PR is about and how this is related to #11375?

@DarkLight1337 Sorry for referencing the incorrect issue number. Please refer to the following issues for
#416
#8323

@DarkLight1337
Copy link
Member

DarkLight1337 commented Jan 2, 2025 •
edited
Loading

There is a related PR #6869, but your solution can sidestep the problem of missing input token IDs (at the cost of being less efficient).

Can you update this PR with the unit tests in #6869 to see whether this solution works correctly?

@Bryce1010 Bryce1010 force-pushed the feature/vllm/add-input-embedding branch 2 times, most recently from 5963444 to f70bbb3 Compare January 6, 2025 11:06
DarkLight1337
DarkLight1337 reviewed Jan 6, 2025
View reviewed changes
vllm/model_executor/models/qwen2.py Outdated
@@ -450,6 +450,7 @@ def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
else:
self.lm_head = ParallelLMHead(config.vocab_size,
config.hidden_size,
True,
Copy link
Member

Choose a reason for hiding this comment

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

I think this will break models that don't have bias weights. Can you read this from the HF config?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Thx, I'll fix it later.

@Bryce1010 Bryce1010 force-pushed the feature/vllm/add-input-embedding branch from f70bbb3 to cdabfaf Compare January 7, 2025 11:41
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@DarkLight1337 DarkLight1337 DarkLight1337 left review comments

@zhuohan123 zhuohan123 Awaiting requested review from zhuohan123 zhuohan123 is a code owner

@youkaichao youkaichao Awaiting requested review from youkaichao youkaichao is a code owner

@alexm-neuralmagic alexm-neuralmagic Awaiting requested review from alexm-neuralmagic alexm-neuralmagic is a code owner

@comaniac comaniac Awaiting requested review from comaniac comaniac is a code owner

@njhill njhill Awaiting requested review from njhill njhill is a code owner

At least 1 approving review is required to merge this pull request.

Assignees
No one assigned
Labels
frontend
Projects
None yet
Milestone
No milestone
2 participants
@Bryce1010 @DarkLight1337