Add vllm#28931

Merged

h-vetinari merged 165 commits into

conda-forge:mainfrom

Jul 29, 2025

Commits on Feb 16, 2025

Add vllm

maresb
authored and
shermansiu
committed
Lints

maresb
authored and
shermansiu
committed
Fix pytorch matchspec

maresb
authored and
shermansiu
committed
Remove extraneous? cuda-nvcc

maresb
authored and
shermansiu
committed
Try adding a bunch of junk from pytorch recipe

maresb
authored and
shermansiu
committed
Revert "Try adding a bunch of junk from pytorch recipe"

maresb
authored and
shermansiu
committed
Add git as host dependency

maresb
authored and
shermansiu
committed
Add swap to conda-forge.yml

maresb
authored and
shermansiu
committed
Bump version

maresb
authored and
shermansiu
committed
Enable verbosity

maresb
authored and
shermansiu
committed
Fix sha256

maresb
authored and
shermansiu
committed
Rework dependencies

maresb
authored and
shermansiu
committed
Check CUDA availability

maresb
authored and
shermansiu
committed
Try some environment variables

maresb
authored and
shermansiu
committed
Remove CUDA_HOME envvar

maresb
authored and
shermansiu
committed
Temporarily revert to fd3cfd7

maresb
authored and
shermansiu
committed
Move build requirements to build section

maresb
authored and
shermansiu
committed
Implement @shermansiu's suggestions

maresb
authored and
shermansiu
committed
Add extra CUDA deps

maresb
authored and
shermansiu
committed
Reformat for recipe.yaml

maresb
authored and
shermansiu
committed
Add zlib

maresb
authored and
shermansiu
committed
Define cuda variables

maresb
authored and
shermansiu
committed
Include xgrammar dependency

maresb
authored and
shermansiu
committed
Don't skip CPU builds

maresb
authored and
shermansiu
committed
Make CUDA dependencies conditional

maresb
authored and
shermansiu
committed
Update to v0.7.2

maresb
authored and
shermansiu
committed
Enforce pytorch ==2.5.1

maresb
authored and
shermansiu
committed
Add libnuma

maresb
authored and
shermansiu
committed
Export build envvars on another line

maresb
authored and
shermansiu
committed
Move vllm to subdirectory

maresb
authored and
shermansiu
committed
Pull in flash-attention as a source

maresb
authored and
shermansiu
committed
Patch CMake for CUDA -> CUDAToolkit

maresb
authored and
shermansiu
committed
Allow adding CMake args for vllm

maresb
authored and
shermansiu
committed
Fix patch path

maresb
authored and
shermansiu
committed
Fix CUDA linking issues.
shermansiu
committed
Only add libnuma for Linux
shermansiu
committed
Ensure sed works with both Linux and Mac.
shermansiu
committed
Don't build for Windows
shermansiu
committed
Indicate that we're using version 2 of the recipe format.
shermansiu
committed
Fix skip selector.
shermansiu
committed
Use v1 config build skipping
shermansiu
committed
Use conda-forge's mistral-common package name
shermansiu
committed
Revert skipping for now
shermansiu
committed
Don't be verbose when building to prevent the Azure logs from crashing
shermansiu
committed
Skip Windows
shermansiu
committed
Fix Windows skipping
shermansiu
committed
Fix build for Linux-cpu
shermansiu
committed
Add megabuild config for Linux
shermansiu
committed
Add macOS-arm64 config
shermansiu
committed
Fix Conda build config
shermansiu
committed
Fix config for Cuda 11.8
shermansiu
committed
Fix recipe selectors.
shermansiu
committed
Fix CUDA recipe selector.
shermansiu
committed
Use Arm 64 image for macOS Arm 64 build
shermansiu
committed
Fix cross-compilation config
shermansiu
committed
Fix Jinja template.
shermansiu
committed
Fix Jinja template
shermansiu
committed
Remove redundant build platform.
shermansiu
committed
Use better Python selector.
shermansiu
committed
Output both vllm and vllm-flash-attn
shermansiu
committed
Drop Python version requirements
shermansiu
committed
Add ignore run exports
shermansiu
committed
Add more license information
shermansiu
committed
Add back MacOS machine to CI config
shermansiu
committed
Fix Cutlass license config
shermansiu
committed
Pin the Cutlass License version
shermansiu
committed
Fix Cutlass license location.
shermansiu
committed
Fix indentation
shermansiu
committed
Use Arm64 image
shermansiu
committed
Only release vllm-flash-attn for Linux CUDA builds
shermansiu
committed
Try a different Mac Arm64 image
shermansiu
committed
Fix the flash-attention import
shermansiu
committed
Ensure that vllm's flash-attention is not installed with the regular flash-attention
shermansiu
committed
Use local copy of Cutlass license
shermansiu
committed
Skip build for macOS x86-64
shermansiu
committed
Publish the Arm-64 version
shermansiu
committed
Fix the provider.
shermansiu
committed

Commits on Feb 17, 2025

Commits on Feb 24, 2025

Commits on Feb 25, 2025

Fix Boolean variable Jinja parsing for rattler-build v0.36.0
shermansiu
committed

Commits on Apr 7, 2025

Commits on Apr 9, 2025

Skip unsupported platforms
shermansiu
committed

Commits on Apr 13, 2025

Fix dependency versions.
shermansiu
committed

Commits on Jul 6, 2025

Commits on Jul 8, 2025

Merge branch 'main' into vllm
shermansiu
authored

Commits on Jul 9, 2025

Commits on Jul 20, 2025

Merge branch 'main' into vllm
shermansiu
authored

Commits on Jul 21, 2025