EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

Clean up import vllm in quantization/__init__.py (#4834)

Browse Source
This commit is contained in:
Lianmin Zheng
2025-03-28 10:34:10 -07:00
committed by GitHub
parent ef9a378a20
commit 74e0ac1dbd
14 changed files with 191 additions and 254 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
scripts/ci_install_dependency.sh
View File

@@ -1,10 +1,8 @@
#!/bin/bash
# Install the dependency in CI.
set -euxo pipefail
# Install the dependency in CI.
# Use repo from environment variable, passed from GitHub Actions
# Use repo from environment variables, passed from GitHub Actions
FLASHINFER_REPO="${FLASHINFER_REPO:-https://flashinfer.ai/whl/cu124/torch2.5/flashinfer-python}"
SCRIPT_DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" && pwd )"
@@ -17,17 +15,12 @@ pip install -e "python[all]" --find-links https://flashinfer.ai/whl/cu124/torch2
rm -rf /root/.cache/flashinfer
# Force reinstall flashinfer and torch_memory_saver
pip install flashinfer_python==0.2.3 --find-links ${FLASHINFER_REPO} --force-reinstall --no-deps
pip install sgl-kernel==0.0.5.post3 --force-reinstall
pip install torch_memory_saver --force-reinstall
pip install transformers==4.50.0 sentence_transformers accelerate==1.4.0 peft pandas datasets
pip install torch_memory_saver
pip install transformers==4.50.0 sentence_transformers accelerate==1.4.0 peft pandas datasets timm
# For compling xgrammar kernels
pip install cuda-python nvidia-cuda-nvrtc-cu12
# For DeepSeek-VL2
pip install timm
pip install sgl-kernel==0.0.5.post3 --force-reinstall
pip uninstall vllm -y || true