[Bugfix] Fix phi3v incorrect image_idx when using async engine #7916
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:
🚀 |
|
/ready |
github-actions
bot
added
the
ready
|
hey @DarkLight1337 @Isotr0py I am getting this error Attempted to assign 457 = 457 multimodal tokens to 757 placeholders 10:36 Does this PR fix this problem? |
Please show the stack trace and the command/code you used. I'm unable to view the image. |
|
hey @DarkLight1337 thanks for the sanppy response the html had issue https://mir-s3-cdn-cf.behance.net/project_modules/max_1200/73fbe271026179.5bb6e7af358b6.jpg [5c964hcpzq] ERROR 09-05 14:03:03 async_llm_engine.py:65] Traceback (most recent call last): |
|
Which version of vLLM are you using? Also which model exactly? |
|
apologies for the delay: [DEBUG][pid:1530] Command to be run: ['/local_disk0/.ephemeral_nfs/envs/pythonEnv-f5e9abdb-8a34-4c63-ad6d-5227ccc7b38f/bin/python', '-m', 'vllm.entrypoints.openai.api_server', '--host', '0.0.0.0', '--port', '9989', '--model', 'microsoft/Phi-3.5-vision-instruct', '--served-model-name', 'microsoft/Phi-3.5-vision-instruct', '--trust-remote-code', '--max-model-len', '12000', '--guided-decoding-backend', 'outlines' vllm is version 0.5.5 pinned |
|
Seems that it is caused by small @stikkireddy Can you add |
|
hmm i dont have chunked prefilling enabled but it consistently fails with that image so i will test that out with chunked prefilling enabled. and also test with 0.6.0. which has a lot of flavors i am looking for. This seemed like the closest issue to mine so i thought i would ask here thanks for the advise so far. I will post a formal issue if this doesnt get resolved otherwise i will post here. |
|
upgrading resolved this issue added this to my test runner job as a test case. thx for the assistance. The upgrade just fixed the problem, i did not need to modify any config. I am trying to offer a OOTB default config and before i add customizations. I am trying to support 3 clouds so adding too many configs makes my life miserable unless i have a easy way to run tests across all clouds and supported models in our weird infra setup. |
FILL IN THE PR DESCRIPTION HERE
FIX #xxxx (link existing issues this PR will resolve)
image_idxbreaks the phi3-vision model server deployment, which doesn't pass prompt to input_processor directly.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!