[Core]: Option To Use Prompt Token Ids Inside Logits Processor

[kezouke]

FIX #4928

UPD (based on comment):

This pull request introduces the option to include prompt token IDs in the logits processor functions.

The updated file vllm/model_executor/layers/logits_processor.py now allows the logits processor to handle either two or three arguments. The choice of arguments passed to the logits processor function is determined by inspecting the function signature. Accordingly, the logits processor will receive either (generated tokens, logits) or (prompt tokens, generated tokens, logits), depending on the expected number of arguments of the function.

---

Old Description:

This pull request introduces a new sampling parameter use_prompt_tokens, which allows the inclusion of prompt token IDs within the logits processor functions. This parameter takes a list of boolean values, one value for each logit processor:

• True : The logits processor function receives the concatenated prompt token IDs and the tokens generated by the model.
• False : The logits processor function receives only the tokens generated by the model.

Additionally, a new validation check has been added to the _verify_args() method in the SamplingParams class to ensure that the logits_processors and use_prompt_tokens lists are of equal length.

---

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, LLM class, 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:

• We adhere to Google Python style guide and Google C++ style guide.
• Pass all linter checks. Please use format.sh to format your code.
• The code need to be well-documented to ensure future contributors can easily understand the code.
• Include sufficient tests to ensure the project to stay correct and robust. This includes both unit tests and integration tests.
• Please add documentation to 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-required and 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:

• After the PR is submitted, the PR will be assigned to a reviewer. Every reviewer will pick up the PRs based on their expertise and availability.
• After the PR is assigned, the reviewer will provide status update every 2-3 days. If the PR is not reviewed within 7 days, please feel free to ping the reviewer or the vLLM team.
• After the review, the reviewer will put an action-required label on the PR if there are changes required. The contributor should address the comments and ping the reviewer to re-review the PR.
• Please respond to all comments within a reasonable time frame. If a comment isn't clear or you disagree with a suggestion, feel free to ask for clarification or discuss the suggestion.

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!

[simon-mo]

Given that we need to align on the length of input as a list, I think another idea would be changing the input attributes to allow three arguments instead of two. You can use https://docs.python.org/3/library/inspect.html#inspect.signature to check the number of arguments.

So in the end we can accept either (past_tokens, logits) or (prompt_tokens, past_tokens, logits).

[kezouke]

Given that we need to align on the length of input as a list, I think another idea would be changing the input attributes to allow three arguments instead of two. You can use https://docs.python.org/3/library/inspect.html#inspect.signature to check the number of arguments.

So in the end we can accept either (past_tokens, logits) or (prompt_tokens, past_tokens, logits).

Hello! Please revise my implementation of function arguments inspection.

Instead of adding a new sampling parameter, I refactored the code by directly examining the logits processor function signature.

New version vllm/model_executor/layers/logits_processor.py enhances the logits processor by enabling it to accept three or two arguments. It utilizes inspect.signature. Depending on the expected arguments for the logits processor function, either (generated tokens, logits) or (prompt tokens, generated tokens, logits) will be passed to the function.

Sorry about the initial confusion regarding your comment.

[simon-mo]

Can you update the type hint in SamplingParams and doc string for logits processors?

[kezouke]

Can you update the type hint in SamplingParams and doc string for logits processors?

Certainly. I apologise for not considering that aspect earlier. Done!