Repository navigation
chore(sync): port vLLM through 255f14599 - #1787
Conversation
Signed-off-by: AlpinDale <alpindale@gmail.com>
360330a to
a9bad29
Compare
Signed-off-by: AlpinDale <alpindale@gmail.com>
Signed-off-by: AlpinDale <alpindale@gmail.com>
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: fea3b663e7
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| from lmcache.integration.vllm.vllm_v1_adapter import ( | ||
| LMCacheConnectorV1Impl as LMCacheConnectorLatestImpl, | ||
| ) |
There was a problem hiding this comment.
When LMCacheConnectorV1 is used with the default use_native=False, this now loads LMCache's vLLM-specific adapter rather than the Aphrodite adapter used previously. Aphrodite does not install or alias the vllm package, so the adapter's vLLM imports and types fail during connector initialization; ordinary LMCache deployments must continue selecting lmcache.integration.aphrodite.aphrodite_v1_adapter or provide an equivalent compatibility adapter.
Useful? React with 👍 / 👎.
| "4", | ||
| ]); | ||
| let err = BenchConfig::from_args(&parse_args(args)).expect_err("should reject the factor"); | ||
| let msg = err.to_string(); |
There was a problem hiding this comment.
This new assertion converts an error directly with to_string(), bypassing the repository's required ToReportString/AsReport formatting convention; the same newly added pattern appears in the timed-trace and protocol tests and should be converted consistently so error context is preserved.
AGENTS.md reference: rust/AGENTS.md:L19-L21
Useful? React with 👍 / 👎.
1 participant
Sync vLLM through 255f14599 while preserving Sonar-specific runtime, sampling, and quantization features.