Set truncate_prompt_tokens in SamplingParams, silently truncating very large prompts and preventing vllm from throwing exception #198
Conversation
…y large prompts and preventing vllm from throwing exception To prevent vllm throwing exceptions like: ```ERROR 08-17 23:32:15 scheduler.py:86] ValueError: The decoder prompt (length 42861) is longer than the maximum model length of 32768. Make sure that `max_model_len` is no smaller than the number of text tokens. ``` `truncate_prompt_tokens=config.max_model_len-1` is used to ensure at least one output token A similar setting was used before modelscope#172, and got removed without an explanation that I could find
There was a problem hiding this comment.
Summary of Changes
Hello @vadimkantorov, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!
This pull request addresses a critical issue where vllm would throw exceptions when processing prompts that exceed the maximum model length. It resolves this by reintroducing a prompt truncation mechanism, specifically by setting truncate_prompt_tokens in SamplingParams. This ensures that overly long prompts are silently truncated, preventing errors and allowing the system to always generate at least one output token, thereby enhancing the stability and reliability of the model serving.
Highlights
- Prompt Handling: Implemented
truncate_prompt_tokensinSamplingParamsto automatically shorten prompts that exceed the model's maximum context length. - Error Prevention: Prevents
ValueErrorexceptions fromvllmthat occur when input prompts are too long, improving the robustness of the system.
Using Gemini Code Assist
The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.
Invoking Gemini
You can request assistance from Gemini at any point in your pull request via creating an issue comment (i.e. comment on the pull request page) using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands.
| Feature | Command | Description |
|---|---|---|
| Code Review | /gemini review |
Performs a code review for the current pull request in its current state. |
| Pull Request Summary | /gemini summary |
Provides a summary of the current pull request in its current state. |
| Comment | @gemini-code-assist | Responds in comments when explicitly tagged, both in issue comments and review comments. |
| Help | /gemini help |
Displays a list of available commands. |
Customization
To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.
Limitations & Feedback
Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.
You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.
Footnotes
-
Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩
There was a problem hiding this comment.
Code Review
This pull request aims to prevent vLLM from throwing an exception on very large prompts by setting truncate_prompt_tokens. While this is a good approach, the current implementation has a flaw that could still lead to exceptions if max_response_tokens is greater than 1. I've provided a suggestion to correctly calculate the truncation length to account for the maximum response size, making the fix more robust.
|
(Previously |
…y large prompts and preventing vllm from throwing exception (modelscope#198) Co-authored-by: Xuchen Pan <[email protected]>
Partially fixing:
max_prompt_lengthandmax_response_length- seems that prompt truncation is not implemented, and this leads to vllm throws an exception:The decoder prompt (length 42861) is longer than the maximum model length of 32768#197To prevent vllm throwing exceptions like:
ERROR 08-17 23:32:15 scheduler.py:86] ValueError: The decoder prompt (length 42861) is longer than the maximum model length of 32768. Make sure that max_model_len is no smaller than the number of text tokens.truncate_prompt_tokens=config.max_model_len-1is used to ensure at least one output tokenA similar setting was used before #172, and got removed without an explanation that I could find
Description
[Please describe the background, purpose, changes made, and how to test this PR]
Checklist
Please check the following items before code is ready to be reviewed.