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.

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

[Model] Add classification Task with Qwen2ForSequenceClassification #9704

Merged
merged 15 commits into from
Oct 26, 2024

Conversation

kakao-kevin-us
Copy link
Contributor

@kakao-kevin-us kakao-kevin-us commented Oct 25, 2024

FILL IN THE PR DESCRIPTION HERE

FIX #8700 (link existing issues this PR will resolve)

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, 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.

Adding or changing kernels

Each custom kernel needs a schema and one or more implementations to be registered with PyTorch.

  • Make sure custom ops are registered following PyTorch guidelines: Custom C++ and CUDA Operators and The Custom Operators Manual
  • Custom operations that return Tensors require meta-functions. Meta-functions should be implemented and registered in python so that dynamic dims can be handled automatically. See above documents for a description of meta-functions.
  • Use torch.libary.opcheck() to test the function registration and meta-function for any registered ops. See tests/kernels for examples.
  • When changing the C++ signature of an existing op, the schema must be updated to reflect the changes.
  • If a new custom type is needed, see the following document: Custom Class Support in PT2.

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!

Key changes:

  • Implement Qwen2ForSequenceClassification model architecture
  • Add model registration for sequence classification variant
  • Include integration tests for the new model variant
  • Update documentation to reflect new sequence classification capability
  • Usecase is similar like [Model] Support Qwen2.5-Math-RM-72B #8896
python -m vllm.entrypoints.openai.api_server \
	--model jason9693/Qwen2.5-1.5B-apeach \
	--trust-remote-code \
	--served-model-name Qwen2.5-1.5B-APEACH \
	--port 8080 \
	--tensor-parallel-size 8 \
	--enforce-eager

Copy link

👋 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.

🚀

@kakao-kevin-us kakao-kevin-us changed the title [Model] Add Qwen2ForSequenceClassification to Qwen [Model] Add classification Task as Qwen2ForSequenceClassification Oct 25, 2024
@kakao-kevin-us kakao-kevin-us changed the title [Model] Add classification Task as Qwen2ForSequenceClassification [Model] Add classification Task with Qwen2ForSequenceClassification Oct 25, 2024
Copy link
Member

@DarkLight1337 DarkLight1337 left a comment

Choose a reason for hiding this comment

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

Thanks for adding this! Some initial comments.

tests/conftest.py Outdated Show resolved Hide resolved
vllm/model_executor/models/qwen2_cls.py Show resolved Hide resolved
tests/conftest.py Outdated Show resolved Hide resolved
@jason9693
Copy link
Contributor

jason9693 commented Oct 26, 2024

@DarkLight1337 All of issues you commented has been resolved.

@DarkLight1337
Copy link
Member

To improve visibility, I also suggest adding a new entry to the Supported Models page of the docs.

@kakao-kevin-us
Copy link
Contributor Author

To improve visibility, I also suggest adding a new entry to the Supported Models page of the docs.

I just added docs

@DarkLight1337
Copy link
Member

It looks like the model test failed. PTAL.

Copy link
Member

@DarkLight1337 DarkLight1337 left a comment

Choose a reason for hiding this comment

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

The model test has passed, thanks for your time!

@DarkLight1337 DarkLight1337 enabled auto-merge (squash) October 26, 2024 14:56
@github-actions github-actions bot added the ready ONLY add when PR is ready to merge/full CI is needed label Oct 26, 2024
@DarkLight1337
Copy link
Member

Please merge from main to fix the CI failures.

auto-merge was automatically disabled October 26, 2024 15:51

Head branch was pushed to by a user without write access

@kakao-kevin-us
Copy link
Contributor Author

kakao-kevin-us commented Oct 26, 2024

@DarkLight1337

Please merge from main to fix the CI failures.

I just rebased on main.
Let's waiting for the test

@DarkLight1337 DarkLight1337 enabled auto-merge (squash) October 26, 2024 15:55
@DarkLight1337 DarkLight1337 merged commit 6650e6a into vllm-project:main Oct 26, 2024
61 checks passed
@corbt
Copy link

corbt commented Oct 27, 2024

Very cool!

Is this compatible with the LoRA adapters functionality in vLLM? I'd love to be able to deploy a base Qwen2ForSequenceClassification and then have many different reward-model variants. The variants would have different LoRA layers and then separate score layers as well.

@kakao-kevin-us
Copy link
Contributor Author

kakao-kevin-us commented Oct 27, 2024

Very cool!

Is this compatible with the LoRA adapters functionality in vLLM? I'd love to be able to deploy a base Qwen2ForSequenceClassification and then have many different reward-model variants. The variants would have different LoRA layers and then separate score layers as well.

It was tested on the CI, but not sure as I'm not tested on my local server.
If it is broken, you can merge the lora weight and run on the vLLM
https://huggingface.co/docs/peft/main/en/conceptual_guides/lora#merge-lora-weights-into-the-base-model

cooleel pushed a commit to cooleel/vllm that referenced this pull request Oct 28, 2024
cooleel pushed a commit to cooleel/vllm that referenced this pull request Oct 28, 2024
FerdinandZhong pushed a commit to FerdinandZhong/vllm that referenced this pull request Oct 29, 2024
rasmith pushed a commit to rasmith/vllm that referenced this pull request Oct 30, 2024
NickLucche pushed a commit to NickLucche/vllm that referenced this pull request Oct 31, 2024
NickLucche pushed a commit to NickLucche/vllm that referenced this pull request Oct 31, 2024
lk-chen pushed a commit to lk-chen/vllm that referenced this pull request Nov 4, 2024
lk-chen pushed a commit to lk-chen/vllm that referenced this pull request Nov 4, 2024
sumitd2 pushed a commit to sumitd2/vllm that referenced this pull request Nov 14, 2024
KuntaiDu pushed a commit to KuntaiDu/vllm that referenced this pull request Nov 20, 2024
mfournioux pushed a commit to mfournioux/vllm that referenced this pull request Nov 20, 2024
tlrmchlsmth pushed a commit to neuralmagic/vllm that referenced this pull request Nov 23, 2024
@tcapelle
Copy link

tcapelle commented Dec 1, 2024

Is this supported in Docker?
I am getting:

ValueError: Model architectures ['Qwen2ForSequenceClassification'] are not supported for now. Supported architectures: ['AquilaModel', 'AquilaForCausalLM', 'ArcticForCausalLM', 'BaiChuanForCausalLM', 'BaichuanForCausalLM', 'BloomForCausalLM', 'CohereForCausalLM', 'DbrxForCausalLM', 'DeciLMForCausalLM', 'DeepseekForCausalLM', 'DeepseekV2ForCausalLM', 'ExaoneForCausalLM', 'FalconForCausalLM', 'GemmaForCausalLM', 'Gemma2ForCausalLM', 'GPT2LMHeadModel', 'GPTBigCodeForCausalLM', 'GPTJForCausalLM', 'GPTNeoXForCausalLM', 'GraniteForCausalLM', 'GraniteMoeForCausalLM', 'InternLMForCausalLM', 'InternLM2ForCausalLM', 'JAISLMHeadModel', 'JambaForCausalLM', 'LlamaForCausalLM', 'LLaMAForCausalLM', 'MambaForCausalLM', 'MistralForCausalLM', 'MixtralForCausalLM', 'QuantMixtralForCausalLM', 'MptForCausalLM', 'MPTForCausalLM', 'MiniCPMForCausalLM', 'MiniCPM3ForCausalLM', 'NemotronForCausalLM', 'OlmoForCausalLM', 'OlmoeForCausalLM', 'OPTForCausalLM', 'OrionForCausalLM', 'PersimmonForCausalLM', 'PhiForCausalLM', 'Phi3ForCausalLM', 'Phi3SmallForCausalLM', 'PhiMoEForCausalLM', 'Qwen2ForCausalLM', 'Qwen2MoeForCausalLM', 'RWForCausalLM', 'StableLMEpochForCausalLM', 'StableLmForCausalLM', 'Starcoder2ForCausalLM', 'SolarForCausalLM', 'XverseForCausalLM', 'BartModel', 'BartForConditionalGeneration', 'Gemma2Model', 'MistralModel', 'Qwen2ForRewardModel', 'Phi3VForCausalLM', 'Blip2ForConditionalGeneration', 'ChameleonForConditionalGeneration', 'ChatGLMModel', 'ChatGLMForConditionalGeneration', 'FuyuForCausalLM', 'InternVLChatModel', 'LlavaForConditionalGeneration', 'LlavaNextForConditionalGeneration', 'LlavaNextVideoForConditionalGeneration', 'LlavaOnevisionForConditionalGeneration', 'MiniCPMV', 'MolmoForCausalLM', 'NVLM_D', 'PaliGemmaForConditionalGeneration', 'PixtralForConditionalGeneration', 'QWenLMHeadModel', 'Qwen2VLForConditionalGeneration', 'UltravoxModel', 'MllamaForConditionalGeneration', 'EAGLEModel', 'MedusaModel', 'MLPSpeculatorPreTrainedModel']

I don't see any SequenceClassification model support...

@DarkLight1337
Copy link
Member

Which version are you using?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
ready ONLY add when PR is ready to merge/full CI is needed
Projects
None yet
Development

Successfully merging this pull request may close these issues.

[Feature]: Support for Seq classification/Reward models
5 participants