feat(vllm-tensorizer): Upgrade vLLM version and Resolve Related Build Compatibility Issues#98

Merged

Eta0 merged 42 commits intomainfrom

jp/testing/update-vllm

Jun 18, 2025

Commits on Jun 6, 2025

Commits on Jun 12, 2025

feat(vllm-tensorizer): Add MAX_JOBS unset logic; remove custom triton build stages
JustinPerlman
committed

Commits on Jun 13, 2025

fix(vllm-builder): Configure CUDA environment variables for vLLM compilation
JustinPerlman
committed
fix(vllm-tensorizer): Update vLLM commit to a newer version for PyTorch 2.7.0 compatibility
JustinPerlman
committed
fix(vllm-tensorizer): Install missing `regex` module for vLLM build metadata
JustinPerlman
committed
feat(vllm-tensorizer): Switch to upstream vLLM for PyTorch 2.7.0 compatibility
JustinPerlman
committed
feat(vllm-tensorizer): Downgrade vLLM to v0.9.0 for PyTorch 2.7.0 compatibility
JustinPerlman
committed
fix(vllm-tensorizer): Apply CMake patch for nvToolsExt linking issue
JustinPerlman
committed
fix(vllm-builder): Simplify find_library call in nvToolsExt CMake patch
JustinPerlman
committed
fix(vllm-tensorizer): Add missing `)`
JustinPerlman
committed
fix(vllm-tensorizer): Remove Cmake patch
JustinPerlman
committed
fix(vllm-tensorizer): Update base image to CUDA 12.8.1 to resolve build issues
JustinPerlman
committed
fix(vllm-tensorizer): Set `MAX_JOBS` to 2 to prevent OOM during Flash Attention compilation
JustinPerlman
committed

Commits on Jun 16, 2025

Commits on Jun 17, 2025