[Minor] Change warning->warning_once in preprocess #26455
Conversation
Signed-off-by: Zhuohan Li <[email protected]>
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request aims to reduce log spam by switching to logger.warning_once. While the intent is good and consistent with other parts of the code, the underlying warning_once implementation is flawed and will cause the application to crash with a TypeError. I've left detailed comments explaining the issue and suggesting a reversion to logger.warning to prevent these crashes. This is a critical issue that should be addressed before merging.
| def get_bos_token_id(self) -> Optional[int]: | ||
| if self.tokenizer is None: | ||
| logger.warning( | ||
| logger.warning_once( |
There was a problem hiding this comment.
While this change aligns with other parts of the file, logger.warning_once will unfortunately cause a runtime TypeError. The underlying implementation in vllm/logger.py uses @lru_cache on a function that takes the logger instance as its first argument. Since logger instances are not hashable, this will cause a crash when this warning is triggered.
This change, intended to reduce log spam, introduces a crash. The *_once logging functionality appears to be broken throughout the codebase and needs to be fixed centrally in vllm/logger.py.
To avoid introducing a crash, I recommend reverting this to logger.warning until the underlying issue with warning_once is resolved.
| logger.warning_once( | |
| logger.warning( |
| def get_eos_token_id(self) -> Optional[int]: | ||
| if self.tokenizer is None: | ||
| logger.warning( | ||
| logger.warning_once( |
There was a problem hiding this comment.
While this change aligns with other parts of the file, logger.warning_once will unfortunately cause a runtime TypeError. The underlying implementation in vllm/logger.py uses @lru_cache on a function that takes the logger instance as its first argument. Since logger instances are not hashable, this will cause a crash when this warning is triggered.
This change, intended to reduce log spam, introduces a crash. The *_once logging functionality appears to be broken throughout the codebase and needs to be fixed centrally in vllm/logger.py.
To avoid introducing a crash, I recommend reverting this to logger.warning until the underlying issue with warning_once is resolved.
| logger.warning_once( | |
| logger.warning( |
github-actions
bot
added
the
ready
…to loader * 'loader' of https://github.com/dsxsteven/vllm_splitPR: (778 commits) [torchao] Add support for ModuleFqnToConfig using regex (vllm-project#26001) Add: Support for multiple hidden layers in Eagle3 (vllm-project#26164) Enable `RMSNorm` substitution for Transformers backend (vllm-project#26353) [Model] Gemma3: Fix GGUF loading and quantization (vllm-project#26189) Bump Flashinfer to v0.4.0 (vllm-project#26326) Update Dockerfile and install runai-model-streamer[gcs] package (vllm-project#26464) [Core] Relax the LoRA max rank (vllm-project#26461) [CI/Build] Fix model nightly tests (vllm-project#26466) [Hybrid]: Decouple Kernel Block Size from KV Page Size (vllm-project#24486) [Core][KVConnector] Propagate all tokens on resumed preemptions (vllm-project#24926) [MM][Doc] Add documentation for configurable mm profiling (vllm-project#26200) [Hardware][AMD] Enable FlexAttention backend on ROCm (vllm-project#26439) [Bugfix] Incorrect another MM data format in vllm bench throughput (vllm-project#26462) [Bugfix] Catch and log invalid token ids in detokenizer #2 (vllm-project#26445) [Minor] Change warning->warning_once in preprocess (vllm-project#26455) [Bugfix] Set the minimum python version for gpt-oss (vllm-project#26392) [Misc] Redact ray runtime env before logging (vllm-project#26302) Separate MLAAttention class from Attention (vllm-project#25103) [Attention] Register FLASHMLA_SPARSE (vllm-project#26441) [Kernels] Modular kernel refactor (vllm-project#24812) ...
Signed-off-by: Zhuohan Li <[email protected]> Signed-off-by: xuebwang-amd <[email protected]>
4 participants
Purpose
Reduce log spamming by changing warning -> warning_once similar to rest of the file.
Essential Elements of an Effective PR Description Checklist
supported_models.mdandexamplesfor a new model.