From 04b52fa8d6f89e430c6d17d21c577c94c1849207 Mon Sep 17 00:00:00 2001 From: Baizhou Zhang Date: Sun, 23 Nov 2025 23:38:36 -0800 Subject: [PATCH] [chore]Upgrade flashinfer to 0.5.3 (#13751) --- docker/Dockerfile | 2 +- python/pyproject.toml | 4 ++-- python/sglang/srt/entrypoints/engine.py | 2 +- 3 files changed, 4 insertions(+), 4 deletions(-) diff --git a/docker/Dockerfile b/docker/Dockerfile index c602cdc74..c7a3f4893 100644 --- a/docker/Dockerfile +++ b/docker/Dockerfile @@ -18,7 +18,7 @@ ARG PIP_DEFAULT_INDEX ARG UBUNTU_MIRROR ARG GITHUB_ARTIFACTORY=github.com ARG INSTALL_FLASHINFER_JIT_CACHE=0 -ARG FLASHINFER_VERSION=0.5.2 +ARG FLASHINFER_VERSION=0.5.3 ENV DEBIAN_FRONTEND=noninteractive \ CUDA_HOME=/usr/local/cuda \ diff --git a/python/pyproject.toml b/python/pyproject.toml index 8e30e8b9c..836bf572c 100755 --- a/python/pyproject.toml +++ b/python/pyproject.toml @@ -26,8 +26,8 @@ dependencies = [ "datasets", "einops", "fastapi", - "flashinfer_python==0.5.2", # keep it aligned with jit-cache version in Dockerfile - "flashinfer_cubin==0.5.2", + "flashinfer_python==0.5.3", # keep it aligned with jit-cache version in Dockerfile + # "flashinfer_cubin==0.5.3", # TODO: Add it back when the cubin is ready: https://github.com/flashinfer-ai/flashinfer/issues/2133 "gguf", "hf_transfer", "huggingface_hub", diff --git a/python/sglang/srt/entrypoints/engine.py b/python/sglang/srt/entrypoints/engine.py index 4e39e98ae..19c4e29a7 100644 --- a/python/sglang/srt/entrypoints/engine.py +++ b/python/sglang/srt/entrypoints/engine.py @@ -727,7 +727,7 @@ def _set_envs_and_config(server_args: ServerArgs): if server_args.attention_backend == "flashinfer": assert_pkg_version( "flashinfer_python", - "0.5.2", + "0.5.3", "Please uninstall the old version and " "reinstall the latest version by following the instructions " "at https://docs.flashinfer.ai/installation.html.",