Bump vllm from 0.6.3.post1 to 0.22.0 in /experiments/agentcompany/openhands#34
Conversation
Bumps [vllm](https://github.com/vllm-project/vllm) from 0.6.3.post1 to 0.22.0. - [Release notes](https://github.com/vllm-project/vllm/releases) - [Changelog](https://github.com/vllm-project/vllm/blob/main/RELEASE.md) - [Commits](vllm-project/vllm@v0.6.3.post1...v0.22.0) --- updated-dependencies: - dependency-name: vllm dependency-version: 0.22.0 dependency-type: direct:production ... Signed-off-by: dependabot[bot] <support@github.com>
dependabot
Bot
added
dependencies
|
Review the following changes in direct dependencies. Learn more about Socket for GitHub.
|
|
Warning Review the following alerts detected in dependencies. According to your organization's Security Policy, it is recommended to resolve "Warn" alerts. Learn more about Socket for GitHub.
|
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
| uvicorn==0.32.0 | ||
| uvloop==0.21.0 | ||
| vllm==0.6.3.post1 | ||
| vllm==0.22.0 |
There was a problem hiding this comment.
vllm==0.22.0 ships against PyTorch 2.11 (that upgrade landed in vllm v0.20.0 as an explicit breaking change). This file still pins torch==2.4.0+cu121 (line 177) and xformers==0.0.27.post2 (line 202), which are built against torch 2.4. Pip will either refuse to install the new vllm wheel due to the torch version mismatch, or the binary CUDA kernels compiled into the vllm wheel will silently crash at import/runtime. All three pins — torch, xformers, and likely triton==3.0.0 — need to be updated together to match what vllm 0.22.0 requires. Beyond torch, compressed-tensors==0.6.0 and transformers==4.46.2 are also likely too old for a 15-major-version bump in vllm.
Prompt To Fix With AI
This is a comment left during a code review.
Path: experiments/agentcompany/openhands/requirements.txt
Line: 192
Comment:
**PyTorch version incompatibility with vllm 0.22.0**
`vllm==0.22.0` ships against PyTorch 2.11 (that upgrade landed in vllm v0.20.0 as an explicit breaking change). This file still pins `torch==2.4.0+cu121` (line 177) and `xformers==0.0.27.post2` (line 202), which are built against torch 2.4. Pip will either refuse to install the new vllm wheel due to the torch version mismatch, or the binary CUDA kernels compiled into the vllm wheel will silently crash at import/runtime. All three pins — `torch`, `xformers`, and likely `triton==3.0.0` — need to be updated together to match what vllm 0.22.0 requires. Beyond torch, `compressed-tensors==0.6.0` and `transformers==4.46.2` are also likely too old for a 15-major-version bump in vllm.
How can I resolve this? If you propose a fix, please make it concise.
Bumps vllm from 0.6.3.post1 to 0.22.0.
Release notes
Sourced from vllm's releases.
... (truncated)
Commits
0b3ba88Revert "[CPU] Experimentally enable Triton and MRV2 (#43225)"799c3af[BugFix] Fix hard-coded timeout for multi-API-server startup (#43768)64e2523[Bugfix] Passrouted_scaling_factorto FlashInfer TRTLLM BF16 MoE (#43769)a147dd0[ROCm][DSV4] Enable Tilelang MHC replacing torch/triton mhc (#43679)0759293[Bugfix][Kernel] TRTLLM NVFP4 MoE chunking (#43599)a930f5aFix RunAI streamer tensor buffer reuse during weight loading (#43464)40cf020Fix early CUDA init (#43791)8c40613[misc] Bump cutedsl version to 4.5.2 (#43745)5ebdf47[Bugfix] Map reasoning_effort to enable_thinking in chat template kwargs (#43...a94cd6d[MRV2][BugFix] Fix KV connector handling in spec decode case (#43719)Dependabot will resolve any conflicts with this PR as long as you don't alter it yourself. You can also trigger a rebase manually by commenting
@dependabot rebase.Dependabot commands and options
You can trigger Dependabot actions by commenting on this PR:
@dependabot rebasewill rebase this PR@dependabot recreatewill recreate this PR, overwriting any edits that have been made to it@dependabot show <dependency name> ignore conditionswill show all of the ignore conditions of the specified dependency@dependabot ignore this major versionwill close this PR and stop Dependabot creating any more for this major version (unless you reopen the PR or upgrade to it yourself)@dependabot ignore this minor versionwill close this PR and stop Dependabot creating any more for this minor version (unless you reopen the PR or upgrade to it yourself)@dependabot ignore this dependencywill close this PR and stop Dependabot creating any more for this dependency (unless you reopen the PR or upgrade to it yourself)You can disable automated security fix PRs for this repo from the Security Alerts page.
Greptile Summary
This dependabot PR bumps
vllmfrom0.6.3.post1to0.22.0in the fully-pinnedexperiments/agentcompany/openhands/requirements.txt. The jump spans ~15 minor vllm versions, and vllm's own release notes call out that v0.20.0 made a PyTorch 2.11 upgrade a breaking environment dependency — this file still pinstorch==2.4.0+cu121.torch==2.4.0+cu121,xformers==0.0.27.post2, andtriton==3.0.0are all built against PyTorch 2.4 and are incompatible with vllm 0.22.0's wheel, which requires PyTorch 2.11.compressed-tensors==0.6.0,transformers==4.46.2,huggingface-hub==0.23.4) are likely too old to be compatible with a vllm 0.22.0 install and would need coordinated updates.Confidence Score: 3/5
Not safe to merge as-is; the vllm wheel will be incompatible with the pinned torch, xformers, and triton versions, breaking installation in this environment.
vllm 0.22.0 ships with CUDA kernels compiled against PyTorch 2.11, but the file still pins torch==2.4.0+cu121 and xformers==0.0.27.post2 (built for torch 2.4). This means the environment will either fail to install or crash at import. The PR changes only one line but requires a coordinated multi-package update to be usable.
experiments/agentcompany/openhands/requirements.txt — the torch, xformers, triton, compressed-tensors, and transformers pins all need to be updated alongside the vllm bump.
Important Files Changed
Flowchart
%%{init: {'theme': 'neutral'}}%% flowchart TD A["vllm==0.22.0\n(requires torch 2.11+)"] -->|depends on| B["torch 2.11+"] A -->|depends on| C["xformers compatible\nwith torch 2.11"] A -->|depends on| D["triton compatible\nwith torch 2.11"] A -->|depends on| E["compressed-tensors\n(new version)"] A -->|depends on| F["transformers\n(newer version)"] G["requirements.txt pins"] --> H["torch==2.4.0+cu121\n❌ incompatible"] G --> I["xformers==0.0.27.post2\n❌ built for torch 2.4"] G --> J["triton==3.0.0\n⚠️ likely incompatible"] G --> K["compressed-tensors==0.6.0\n⚠️ likely too old"] G --> L["transformers==4.46.2\n⚠️ likely too old"] H -->|conflicts with| B I -->|conflicts with| C J -->|may conflict with| DPrompt To Fix All With AI
Reviews (1): Last reviewed commit: "Bump vllm in /experiments/agentcompany/o..." | Re-trigger Greptile