chore: bump v0.3.6.post3 (#2259)

Makefile

-.PHONY: check-deps install-deps format
+.PHONY: check-deps install-deps format update
 
 check-deps:
 	@command -v isort >/dev/null 2>&1 || (echo "Installing isort..." && pip install isort)
@@ -10,3 +10,29 @@ install-deps:
 format: check-deps
 	@echo "Formatting modified Python files..."
 	git diff --name-only --diff-filter=M | grep '\.py$$' | xargs -I {} sh -c 'isort {} && black {}'
+
+FILES_TO_UPDATE = docker/Dockerfile.rocm \
+                 python/pyproject.toml \
+                 python/sglang/version.py \
+                 docs/developer/setup_github_runner.md \
+                 docs/start/install.md
+
+update:
+	@if [ -z "$(filter-out $@,$(MAKECMDGOALS))" ]; then \
+		echo "Version required. Usage: make update <new_version>"; \
+		exit 1; \
+	fi
+	@OLD_VERSION=$$(grep "version" python/sglang/version.py | cut -d '"' -f2); \
+	NEW_VERSION=$(filter-out $@,$(MAKECMDGOALS)); \
+	echo "Updating version from $$OLD_VERSION to $$NEW_VERSION"; \
+	for file in $(FILES_TO_UPDATE); do \
+		if [ "$(shell uname)" = "Darwin" ]; then \
+			sed -i '' -e "s/$$OLD_VERSION/$$NEW_VERSION/g" $$file; \
+		else \
+			sed -i -e "s/$$OLD_VERSION/$$NEW_VERSION/g" $$file; \
+		fi \
+	done; \
+	echo "Version update complete"
+
+%:
+	@:

docker/Dockerfile

@@ -23,31 +23,32 @@ RUN pip3 install datamodel_code_generator
 
 WORKDIR /sgl-workspace
 
+ARG CUDA_VERSION
 RUN python3 -m pip install --upgrade pip setuptools wheel html5lib six \
     && git clone --depth=1 https://github.com/sgl-project/sglang.git \
     && cd sglang \
     && if [ "$BUILD_TYPE" = "srt" ]; then \
-         python3 -m pip --no-cache-dir install -e "python[srt]"; \
+         if [ "$CUDA_VERSION" = "12.1.1" ]; then \
+           python3 -m pip --no-cache-dir install -e "python[srt]" --find-links https://flashinfer.ai/whl/cu121/torch2.4/flashinfer/; \
+         elif [ "$CUDA_VERSION" = "12.4.1" ]; then \
+           python3 -m pip --no-cache-dir install -e "python[srt]" --find-links https://flashinfer.ai/whl/cu124/torch2.4/flashinfer/; \
+         elif [ "$CUDA_VERSION" = "11.8.0" ]; then \
+           python3 -m pip --no-cache-dir install -e "python[srt]" --find-links https://flashinfer.ai/whl/cu118/torch2.4/flashinfer/; \
+         else \
+           echo "Unsupported CUDA version: $CUDA_VERSION" && exit 1; \
+         fi; \
        else \
-         python3 -m pip --no-cache-dir install -e "python[all]"; \
+         if [ "$CUDA_VERSION" = "12.1.1" ]; then \
+           python3 -m pip --no-cache-dir install -e "python[all]" --find-links https://flashinfer.ai/whl/cu121/torch2.4/flashinfer/; \
+         elif [ "$CUDA_VERSION" = "12.4.1" ]; then \
+           python3 -m pip --no-cache-dir install -e "python[all]" --find-links https://flashinfer.ai/whl/cu124/torch2.4/flashinfer/; \
+         elif [ "$CUDA_VERSION" = "11.8.0" ]; then \
+           python3 -m pip --no-cache-dir install -e "python[all]" --find-links https://flashinfer.ai/whl/cu118/torch2.4/flashinfer/; \
+         else \
+           echo "Unsupported CUDA version: $CUDA_VERSION" && exit 1; \
+         fi; \
        fi
 
-ARG CUDA_VERSION
-RUN if [ "$CUDA_VERSION" = "12.1.1" ]; then \
-        export CUDA_IDENTIFIER=cu121 && \
-        python3 -m pip --no-cache-dir install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/; \
-    elif [ "$CUDA_VERSION" = "12.4.1" ]; then \
-        export CUDA_IDENTIFIER=cu124 && \
-        python3 -m pip --no-cache-dir install flashinfer -i https://flashinfer.ai/whl/cu124/torch2.4/; \
-    elif [ "$CUDA_VERSION" = "11.8.0" ]; then \
-        export CUDA_IDENTIFIER=cu118 && \
-        python3 -m pip install torch==2.4.0 --index-url https://download.pytorch.org/whl/cu118 && \
-        python3 -m pip --no-cache-dir install flashinfer -i https://flashinfer.ai/whl/cu118/torch2.4/; \
-    else \
-        echo "Unsupported CUDA version: $CUDA_VERSION" && exit 1; \
-    fi
-
 RUN python3 -m pip cache purge
 
-
 ENV DEBIAN_FRONTEND=interactive

docker/Dockerfile.rocm

 # Usage (to build SGLang ROCm docker image):
-#   docker build --build-arg SGL_BRANCH=v0.3.6.post2 -t v0.3.6.post2-rocm620 -f Dockerfile.rocm .
+#   docker build --build-arg SGL_BRANCH=v0.3.6.post3 -t v0.3.6.post3-rocm620 -f Dockerfile.rocm .
 
 # default base image
 ARG BASE_IMAGE="rocm/vllm-dev:20241022"

docs/developer/setup_github_runner.md

@@ -11,9 +11,9 @@ docker pull nvidia/cuda:12.1.1-devel-ubuntu22.04
 # Nvidia
 docker run --shm-size 128g -it -v /tmp/huggingface:/hf_home --gpus all nvidia/cuda:12.1.1-devel-ubuntu22.04 /bin/bash
 # AMD
-docker run --rm --device=/dev/kfd --device=/dev/dri --group-add video --shm-size 128g -it -v /tmp/huggingface:/hf_home lmsysorg/sglang:v0.3.6.post2-rocm620 /bin/bash
+docker run --rm --device=/dev/kfd --device=/dev/dri --group-add video --shm-size 128g -it -v /tmp/huggingface:/hf_home lmsysorg/sglang:v0.3.6.post3-rocm620 /bin/bash
 # AMD just the last 2 GPUs
-docker run --rm --device=/dev/kfd --device=/dev/dri/renderD176 --device=/dev/dri/renderD184 --group-add video --shm-size 128g -it -v /tmp/huggingface:/hf_home lmsysorg/sglang:v0.3.6.post2-rocm620 /bin/bash
+docker run --rm --device=/dev/kfd --device=/dev/dri/renderD176 --device=/dev/dri/renderD184 --group-add video --shm-size 128g -it -v /tmp/huggingface:/hf_home lmsysorg/sglang:v0.3.6.post3-rocm620 /bin/bash
 ```
 
 ### Step 2: Configure the runner by `config.sh`

docs/start/install.md

@@ -5,10 +5,7 @@ You can install SGLang using any of the methods below.
 ## Method 1: With pip
 ```
 pip install --upgrade pip
-pip install "sglang[all]"
-
-# Install FlashInfer accelerated kernels (CUDA only for now)
-pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/
+pip install "sglang[all]" --find-links https://flashinfer.ai/whl/cu121/torch2.4/flashinfer/
 ```
 
 Note: Please check the [FlashInfer installation doc](https://docs.flashinfer.ai/installation.html) to install the proper version according to your PyTorch and CUDA versions.
@@ -16,14 +13,11 @@ Note: Please check the [FlashInfer installation doc](https://docs.flashinfer.ai/
 ## Method 2: From source
 ```
 # Use the last release branch
-git clone -b v0.3.6.post2 https://github.com/sgl-project/sglang.git
+git clone -b v0.3.6.post3 https://github.com/sgl-project/sglang.git
 cd sglang
 
 pip install --upgrade pip
-pip install -e "python[all]"
-
-# Install FlashInfer accelerated kernels (CUDA only for now)
-pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/
+pip install -e "python[all]" --find-links https://flashinfer.ai/whl/cu121/torch2.4/flashinfer/
 ```
 
 Note: Please check the [FlashInfer installation doc](https://docs.flashinfer.ai/installation.html) to install the proper version according to your PyTorch and CUDA versions.
@@ -32,7 +26,7 @@ Note: To AMD ROCm system with Instinct/MI GPUs, do following instead:
 
 ```
 # Use the last release branch
-git clone -b v0.3.6.post2 https://github.com/sgl-project/sglang.git
+git clone -b v0.3.6.post3 https://github.com/sgl-project/sglang.git
 cd sglang
 
 pip install --upgrade pip
@@ -57,7 +51,7 @@ docker run --gpus all \
 Note: To AMD ROCm system with Instinct/MI GPUs, it is recommended to use `docker/Dockerfile.rocm` to build images, example and usage as below:
 
 ```bash
-docker build --build-arg SGL_BRANCH=v0.3.6.post2 -t v0.3.6.post2-rocm620 -f Dockerfile.rocm .
+docker build --build-arg SGL_BRANCH=v0.3.6.post3 -t v0.3.6.post3-rocm620 -f Dockerfile.rocm .
 
 alias drun='docker run -it --rm --network=host --device=/dev/kfd --device=/dev/dri --ipc=host \
     --shm-size 16G --group-add video --cap-add=SYS_PTRACE --security-opt seccomp=unconfined \
@@ -66,11 +60,11 @@ alias drun='docker run -it --rm --network=host --device=/dev/kfd --device=/dev/d
 drun -p 30000:30000 \
     -v ~/.cache/huggingface:/root/.cache/huggingface \
     --env "HF_TOKEN=<secret>" \
-    v0.3.6.post2-rocm620 \
+    v0.3.6.post3-rocm620 \
     python3 -m sglang.launch_server --model-path meta-llama/Llama-3.1-8B-Instruct --host 0.0.0.0 --port 30000
 
 # Till flashinfer backend available, --attention-backend triton --sampling-backend pytorch are set by default
-drun v0.3.6.post2-rocm620 python3 -m sglang.bench_one_batch --batch-size 32 --input 1024 --output 128 --model amd/Meta-Llama-3.1-8B-Instruct-FP8-KV --tp 8 --quantization fp8
+drun v0.3.6.post3-rocm620 python3 -m sglang.bench_one_batch --batch-size 32 --input 1024 --output 128 --model amd/Meta-Llama-3.1-8B-Instruct-FP8-KV --tp 8 --quantization fp8
 ```
 
 ## Method 4: Using docker compose

python/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "sglang"
-version = "0.3.6.post2"
+version = "0.3.6.post3"
 description = "SGLang is yet another fast serving framework for large language models and vision language models."
 readme = "README.md"
 requires-python = ">=3.8"
@@ -23,7 +23,7 @@ runtime_common = ["aiohttp", "decord", "fastapi",
     "psutil", "pydantic", "python-multipart",
     "pyzmq>=25.1.2", "torchao", "uvicorn", "uvloop",
     "xgrammar>=0.1.4"]
-srt = ["sglang[runtime_common]", "torch", "vllm>=0.6.3.post1", "cuda-python"]
+srt = ["sglang[runtime_common]", "torch", "vllm>=0.6.3.post1", "cuda-python", "flashinfer>=0.1.6"]
 
 # HIP (Heterogeneous-computing Interface for Portability) for AMD
 # => base docker rocm/vllm-dev:20241022, not from public vllm whl

python/sglang/version.py

-__version__ = "0.3.6.post2"
+__version__ = "0.3.6.post3"

scripts/ci_install_dependency.sh

@@ -3,9 +3,8 @@ Install the dependency in CI.
 """
 
 pip install --upgrade pip
-pip install -e "python[all]"
+pip install -e "python[all]" --find-links https://flashinfer.ai/whl/cu121/torch2.4/flashinfer/
 pip install transformers==4.45.2 sentence_transformers accelerate peft
-pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/ --force-reinstall
 # for compling eagle kernels
 pip install cutex
 # for compling xgrammar kernels