npx claudepluginhub opendatahub-io/ai-helpers --plugin odh-ai-helpersThis skill is limited to using the following tools:
Compare vllm requirements files **and Dockerfiles** between versions to identify dependency changes, providing intelligent analysis for AIPCC package onboarding workflows.
Deploys vLLM inference server using Docker (pre-built images or build-from-source) with NVIDIA GPU support and OpenAI-compatible API.
Deep-dives into ML/AI topics by fetching official docs and GitHub sources via KB or web tools, for explaining concepts, comparing approaches, or surveying frameworks like 'how does X work?' or 'X vs Y'.
Detects performance regressions across git versions by comparing benchmarks for latency (p50/p95/p99), throughput, memory, CPU/I/O, with statistical tests and reports.
Share bugs, ideas, or general feedback.
Compare vllm requirements files and Dockerfiles between versions to identify dependency changes, providing intelligent analysis for AIPCC package onboarding workflows.
For accelerator builds (ROCm, CUDA, TPU, XPU), Dockerfile comparison is critical because they specify exact commits/branches for dependencies built from source (PyTorch, Triton, Flash Attention, etc.) - information not available in requirements files.
Note: Each variant compares only the files that actually exist in the vllm repository. For example, CUDA has cuda.txt but no cuda-build.txt, while ROCm has both rocm.txt and rocm-build.txt.
Executable Script - Runs a Python script to fetch and compare requirements files from the vllm GitHub repository.
Use this skill when you need to:
Note: Variant comparisons include BOTH runtime and build dependencies to ensure complete coverage.
./scripts/compare_reqs.py <version1> <version2> <variant|file> [--pretty]
v0.13.0, v0.14.0rc1)rocm, cuda, cpu, tpu, xpu (auto-includes runtime + build requirements + Dockerfiles)rocm-build.txt, common.txt, docker/Dockerfile.rocm, etc.# Compare ROCm runtime + build requirements + Dockerfiles
./scripts/compare_reqs.py v0.13.0 v0.14.0 rocm
# Compare CUDA runtime + build requirements + Dockerfiles
./scripts/compare_reqs.py v0.13.0 v0.14.0 cuda
# Compare specific file only
./scripts/compare_reqs.py v0.13.0 v0.14.0 common.txt
./scripts/compare_reqs.py v0.13.0 v0.14.0 rocm-build.txt
./scripts/compare_reqs.py v0.13.0 v0.14.0 docker/Dockerfile.rocm_base
# All variants (based on what files actually exist in vllm repo)
./scripts/compare_reqs.py v0.13.0 v0.14.0 rocm # common.txt + rocm.txt + rocm-build.txt + Dockerfiles
./scripts/compare_reqs.py v0.13.0 v0.14.0 cuda # common.txt + cuda.txt + Dockerfile
./scripts/compare_reqs.py v0.13.0 v0.14.0 cpu # common.txt + cpu.txt + cpu-build.txt + Dockerfile.cpu
./scripts/compare_reqs.py v0.13.0 v0.14.0 tpu # common.txt + tpu.txt + Dockerfile.tpu
./scripts/compare_reqs.py v0.13.0 v0.14.0 xpu # common.txt + xpu.txt + Dockerfile.xpu (Intel GPU)
The script provides clean, categorized output with a summary table followed by detailed changes:
=== Comparing rocm variant (build + Dockerfiles): v0.13.0 -> v0.14.0rc1 ===
๐ Change Summary Table:
File Package Old Version New Version Type
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
common.txt protobuf - >= 6.30.0 Added
common.txt grpcio - >=1.76.0 Added
rocm-build.txt torch ==2.9.0 ==2.9.1 Changed
rocm-build.txt triton ==3.5.0 ==3.5.1 Changed
docker/Dockerfil... PYTORCH_BRANCH=1c57644d Changed
docker/Dockerfil... MORI_BRANCH=2d02c6a9 - Added
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
๐ common.txt
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
๐ฆ Changed:
xgrammar == 0.1.27 โ xgrammar == 0.1.29
mistral_common[image] >= 1.8.5 โ mistral_common[image] >= 1.8.8
โ Added:
protobuf >= 6.30.0 # Required by LlamaTokenizer, gRPC.
grpcio>=1.76.0
โ Removed:
scipy # Required for phi-4-multimodal-instruct
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
๐ rocm-build.txt
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
๐ฆ Changed:
torch==2.9.0 โ torch==2.9.1
triton==3.5.0 โ triton==3.5.1
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
๐ณ docker/Dockerfile.rocm_base
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
๐ฆ Changed:
BASE_IMAGE=rocm/dev-ubuntu-22.04:7.1-complete โ BASE_IMAGE=rocm/dev-ubuntu-22.04:7.0-complete
PYTORCH_BRANCH=1c57644d โ PYTORCH_BRANCH=89075173
AITER_BRANCH=59bd8ff2 โ AITER_BRANCH=6af8b687
โ Added:
MORI_BRANCH=2d02c6a9
MORI_REPO=https://github.com/ROCm/mori.git
RIXL_BRANCH=50d63d94
RIXL_REPO=https://github.com/vcave/RIXL.git
After running the script and displaying the output, you must provide intelligent impact analysis:
## Summary of Changes: vllm v0.13.0 โ v0.14.0rc1
### Impact Level: Medium
### AIPCC Wheels Builder Impact:
#### โ
No Action Required:
- ROCm version remains at 6.4
- PyTorch ecosystem stable (minor patch updates only)
#### โ ๏ธ Action Required:
**New packages to onboard:**
1. **`grpcio>=1.76.0`** (NEW)
- Purpose: gRPC support
- Complexity: Compiled extension (C++)
- PyPI: https://pypi.org/project/grpcio/
- Action: Onboard grpcio 1.76.0+
**Version updates:**
- torch: 2.9.0 โ 2.9.1 (patch - safe)
- triton: 3.5.0 โ 3.5.1 (patch - safe)
**Removed dependencies:**
- scipy: Was required for phi-4-multimodal-instruct
- Risk: Low (model-specific, may not be widely used)
#### Next Steps:
1. Onboard grpcio>=1.76.0 to wheels builder
2. Verify torch 2.9.1 and triton 3.5.1 availability
3. Test vllm 0.14.0rc1 build with new dependencies
4. Validate phi-4 model functionality without scipy
**Context:** This release adds gRPC support and updates to PyTorch 2.9.1.
When onboarding a new vllm version:
When builds fail:
When planning releases:
The script gracefully handles: