[Feature][Frontend]: Continued stream_options implementation also in CompletionRequest
#5319
Conversation
Add stream_options validation in CompletionRequest
…usage` field based on `stream_options.include_usage`. - Enhanced the token-by-token and finish responses to conditionally include `usage` field if `stream_options.include_usage` is set. - Added a final usage statistics message if `stream_options.include_usage` is set, including prompt tokens and completion tokens.
- stream=True, stream_options=None
- stream=True, stream_options={"include_usage": True}
- stream=True, stream_options={"include_usage": False}
- stream=False, stream_options={"include_usage": None}
- stream=False, stream_options={"include_usage": False}
- stream=False, stream_options={"include_usage": True}
format.sh.
|
Unlike Chat Completions API, Completions API doesn't have a response format specific to streaming. Based on the documentation, it seems that |
|
Will make the needed adjustments later today! |
Actually it seems completions works just the same regarding this: `include_usage boolean If set, an additional chunk will be streamed before the data: [DONE] message. The usage field on this chunk shows the token usage statistics for the entire request, and the choices field will always be an empty array. All other chunks will also include a usage field, but with a null value.` |
Looks like you're right. I focused too much on the return value. In that case, I'll approve the PR once you have fixed the test cases and resolved the merge conflict. |
Noted by DrakLIght there was two issues: 1. Notation of `@pytest.mark.asyncio` on the test function. 2. Checking on chunk usage on a non exisiting variable.
Should have done that on prev commit TBH
-- Removed redundent StreamOptions. -- Formater.sh
|
I've changed my mind regarding |
-- Added parametrize in completion stream options. -- Revised streaming tests as the usage is no longer needed to be asserted.
-- single_usage is no longer needed inside test completion streaming.
-- Resolved concerns raised by DarkLight (Mistake related to client.chat.completions.creat) -- Resolved issue related to MODEL_NAME,
-- Redundent test removed (stream=True, stream_options=None)
-- Not found.
|
Try placing the stream options tests before the embeddings tests. There might be conflicts between the pytest fixtures. |
- Moved stream options tests before embeddings tests to address conflicts between pytest fixtures. - This change is in response to a suggestion from DarkLight1337. - Adjustments made to ensure test suite runs without errors.
- Moved stream options tests before embeddings tests to address conflicts between pytest fixtures. - This change is in response to a suggestion from DarkLight1337. - Adjustments made to ensure test suite runs without errors. -- Formated code
- **Stream with `include_usage: False`**:
- Added assertions to ensure no chunk contains the `usage` key.
- **Stream with `include_usage: True`**:
- Modified test logic to verify that every chunk has `usage` as `None` except for the last chunk, which should have `usage` populated.
- **Stream=False configurations**:
- Added tests to verify that using `stream_options: {"include_usage": None}`, `{"include_usage": False}`, and `{"include_usage": True}` raises a `BadRequestError`.
- Removed redundant test for `stream=False` with `stream_options: {"include_usage": False}` as it overlaps with the error condition checks.
…tribute - **Stream with `include_usage: False`**: - Updated tests to assert that the `usage` attribute is `None` instead of checking its absence in the chunk dictionary. This aligns with the observed behavior where `usage` is present but set to `None`.
- Incorrect indent caused an empty `choices` list after each generated `choices`. - Moved the final `usage` creation step back one indent level to fix this.
- Incorrect indent caused double sending of a chuck resulting in a server crush. - Moved the final `usage` creation step back twp indent levels to fix this.
|
The entrypoints tests are finally passing. Thanks again for your efforts! |
Yeah, had some minor issues there. Thank you for the help!! |
…n CompletionRequest (vllm-project#5319)
* upstream/main: (126 commits) [Bugfix][Frontend] Cleanup "fix chat logprobs" (vllm-project#5026) [Bugfix] OpenAI entrypoint limits logprobs while ignoring server defined --max-logprobs (vllm-project#5312) [Misc] Various simplifications and typing fixes (vllm-project#5368) [ci] Fix Buildkite agent path (vllm-project#5392) [Doc] Add documentation for FP8 W8A8 (vllm-project#5388) Bump version to v0.5.0 (vllm-project#5384) [Docs] Alphabetically sort sponsors (vllm-project#5386) [Docs] Add Docs on Limitations of VLM Support (vllm-project#5383) [ci] Mount buildkite agent on Docker container to upload benchmark results (vllm-project#5330) [ci] Use small_cpu_queue for doc build (vllm-project#5331) [Bugfix] Fix LLaVA-NeXT (vllm-project#5380) [Feature][Frontend]: Continued `stream_options` implementation also in CompletionRequest (vllm-project#5319) [Model] Initial support for LLaVA-NeXT (vllm-project#4199) [Misc] Improve error message when LoRA parsing fails (vllm-project#5194) [misc][typo] fix typo (vllm-project#5372) [Frontend][Misc] Enforce Pixel Values as Input Type for VLMs in API Server (vllm-project#5374) [Misc] Update to comply with the new `compressed-tensors` config (vllm-project#5350) [Bugfix] Fix KeyError: 1 When Using LoRA adapters (vllm-project#5164) [Kernel][Misc] Use TORCH_LIBRARY instead of PYBIND11_MODULE for custom ops (vllm-project#5047) [mis][ci/test] fix flaky test in test_sharded_state_loader.py (vllm-project#5361) ...
…n CompletionRequest (vllm-project#5319)
…n CompletionRequest (vllm-project#5319)
…n CompletionRequest (vllm-project#5319)
…n CompletionRequest (vllm-project#5319)
…n CompletionRequest (vllm-project#5319)
This PR introduces support for the
stream_optionparameter in the CompletionRequest class.Continuing the previous PR: #5135
FIX #4967(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,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!