[Bugfix] spec decode handle None entries in topk args in create_sequence_group_output #7232
Conversation
|
👋 Hi! Thank you for contributing to the vLLM project. Once the PR is approved and ready to go, please make sure to run full CI as it is required to merge (or just use auto-merge). To run full CI, you can do one of these:
🚀 |
tjohnson31415
changed the title
tjohnson31415
changed the title
tjohnson31415
force-pushed
the
spec-decoding-logprobs
branch
from
e139f24 to
7d267e0
Compare
Signed-off-by: Travis Johnson <[email protected]>
Signed-off-by: Travis Johnson <[email protected]>
tjohnson31415
force-pushed
the
spec-decoding-logprobs
branch
from
7d267e0 to
dcba1da
Compare
|
Coming back to this PR: i have also added a test to check the behavior with logprobs for spec decoding disabled, which would have caught this bug. |
…nce_group_output (vllm-project#7232) Signed-off-by: Travis Johnson <[email protected]>
…nce_group_output (vllm-project#7232) Signed-off-by: Travis Johnson <[email protected]> Signed-off-by: Alvant <[email protected]>
…nce_group_output (vllm-project#7232) Signed-off-by: Travis Johnson <[email protected]>
Currently, requesting logprobs for generated tokens when using speculative decoding is broken. Sending a request for logprobs causes the server to crash with:
I traced this down to there being an entry in the
logprobsof the sampler outputs that hasNonefor the token id. The PR #6485, optimized speculative decoding by removing unnecessary operations on logprobs. This is configured with thedisable_logprobs_during_spec_decodingflag, which is enabled by default (i.e. logprob computations are skipped by default). Instead of returning real logprobs,_create_dummy_logprob_listscreates dummy outputs withNoneentries for thetop_ktoken ids and logprobs which leads to the error above in later processing. The change in this PR is to ignore theseNoneentries increate_sequence_group_output.FIX #6967
BEFORE SUBMITTING, PLEASE READ THE CHECKLIST BELOW AND FILL IN THE DESCRIPTION ABOVE
PR Checklist (Click to Expand)
Thank you for your contribution to vLLM! Before submitting the pull request, please ensure the PR meets the following criteria. This helps vLLM maintain the code quality and improve the efficiency of the review process.
PR Title and Classification
Only specific types of PRs will be reviewed. The PR title is prefixed appropriately to indicate the type of change. Please use one of the following:
[Bugfix]for bug fixes.[CI/Build]for build or continuous integration improvements.[Doc]for documentation fixes and improvements.[Model]for adding a new model or improving an existing model. Model name should appear in the title.[Frontend]For changes on the vLLM frontend (e.g., OpenAI API server,LLMclass, etc.)[Kernel]for changes affecting CUDA kernels or other compute kernels.[Core]for changes in the core vLLM logic (e.g.,LLMEngine,AsyncLLMEngine,Scheduler, etc.)[Hardware][Vendor]for hardware-specific changes. Vendor name should appear in the prefix (e.g.,[Hardware][AMD]).[Misc]for PRs that do not fit the above categories. Please use this sparingly.Note: If the PR spans more than one category, please include all relevant prefixes.
Code Quality
The PR need to meet the following code quality standards:
format.shto format your code.docs/source/if the PR modifies the user-facing behaviors of vLLM. It helps vLLM user understand and utilize the new features or changes.Notes for Large Changes
Please keep the changes as concise as possible. For major architectural changes (>500 LOC excluding kernel/data/config/test), we would expect a GitHub issue (RFC) discussing the technical design and justification. Otherwise, we will tag it with
rfc-requiredand might not go through the PR.What to Expect for the Reviews
The goal of the vLLM team is to be a transparent reviewing machine. We would like to make the review process transparent and efficient and make sure no contributor feel confused or frustrated. However, the vLLM team is small, so we need to prioritize some PRs over others. Here is what you can expect from the review process:
action-requiredlabel on the PR if there are changes required. The contributor should address the comments and ping the reviewer to re-review the PR.Thank You
Finally, thank you for taking the time to read these guidelines and for your interest in contributing to vLLM. Your contributions make vLLM a great tool for everyone!