Update ci workflows (#1804)

.github/workflows/deploy-docs.yml

@@ -1,20 +1,15 @@
 name: Build Documentation
 
 on:
-  push:
-    branches: [ main ]
-  pull_request:
-    branches: [ main ]
   workflow_dispatch:
 
 jobs:
   execute-notebooks:
     runs-on: 1-gpu-runner
-    if: github.event_name == 'pull_request' || github.event_name == 'workflow_dispatch'
+    if: github.repository == 'sgl-project/sglang' || github.event_name == 'pull_request'
     steps:
-      - uses: actions/checkout@v3
-        with:
-          fetch-depth: 0
+      - name: Checkout code
+        uses: actions/checkout@v3
 
       - name: Set up Python
         uses: actions/setup-python@v4
@@ -23,22 +18,14 @@ jobs:
 
       - name: Install dependencies
         run: |
-          pip install --upgrade pip
-          pip install -e "python[all]"
+          bash scripts/ci_install_dependency.sh
           pip install -r docs/requirements.txt
-          pip install nbconvert jupyter_client ipykernel ipywidgets matplotlib
-          pip install transformers==4.45.2
-          pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/ --force-reinstall
 
       - name: Setup Jupyter Kernel
         run: |
           python -m ipykernel install --user --name python3 --display-name "Python 3"
 
       - name: Execute notebooks
-        env:
-          HF_HOME: /hf_home
-          SGLANG_IS_IN_CI: true
-          CUDA_VISIBLE_DEVICES: 0
         run: |
           cd docs/en
           for nb in *.ipynb; do
@@ -54,34 +41,18 @@ jobs:
     if: github.event_name == 'push' && github.ref == 'refs/heads/main'
     runs-on: 1-gpu-runner
     steps:
-      - uses: actions/checkout@v3
-        with:
-          fetch-depth: 0
+      - name: Checkout code
+        uses: actions/checkout@v3
 
       - name: Set up Python
         uses: actions/setup-python@v4
         with:
           python-version: '3.9'
 
-      - name: Cache Python dependencies
-        uses: actions/cache@v3
-        with:
-          path: ~/.cache/pip
-          key: ${{ runner.os }}-pip-${{ hashFiles('**/requirements.txt') }}
-          restore-keys: |
-            ${{ runner.os }}-pip-
-
       - name: Install dependencies
         run: |
-          pip install --upgrade pip
-          pip install -e "python[all]"
+          bash scripts/ci_install_dependency.sh
           pip install -r docs/requirements.txt
-          pip install nbconvert jupyter_client ipykernel ipywidgets matplotlib
-          pip install transformers==4.45.2
-          pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/ --force-reinstall
-
-      - name: Install Pandoc
-        run: |
           apt-get update
           apt-get install -y pandoc
 

.github/workflows/lint.yml

@@ -8,10 +8,10 @@ jobs:
     steps:
       - uses: actions/checkout@v2
 
-      - name: Set up Python 3.9
-        uses: actions/setup-python@v2
+      - name: Set up Python
+        uses: actions/setup-python@v4
         with:
-          python-version: 3.9
+          python-version: '3.9'
 
       - name: Install pre-commit hook
         run: |

.github/workflows/nightly-eval.yml

@@ -24,9 +24,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          pip install --upgrade pip
-          pip install -e "python[all]"
-          pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/ --force-reinstall
+          bash scripts/ci_install_dependency.sh
 
       - name: Nightly gsm8k Accuracy
         timeout-minutes: 60

.github/workflows/pr-test.yml

@@ -27,10 +27,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          pip install --upgrade pip
-          pip install -e "python[dev]"
-          pip install transformers==4.45.2
-          pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/ --force-reinstall
+          bash scripts/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 10
@@ -47,10 +44,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          pip install --upgrade pip
-          pip install -e "python[dev]"
-          pip install transformers==4.45.2
-          pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/ --force-reinstall
+          bash scripts/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 20
@@ -67,10 +61,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          pip install --upgrade pip
-          pip install -e "python[dev]"
-          pip install transformers==4.45.2
-          pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/ --force-reinstall
+          bash scripts/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 20
@@ -87,10 +78,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          pip install --upgrade pip
-          pip install -e "python[dev]"
-          pip install transformers==4.45.2
-          pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/ --force-reinstall
+          bash scripts/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 20
@@ -107,10 +95,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          pip install --upgrade pip
-          pip install -e "python[dev]"
-          pip install transformers==4.45.2
-          pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/ --force-reinstall
+          bash scripts/ci_install_dependency.sh
 
       - name: Run test
         timeout-minutes: 20
@@ -127,10 +112,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          pip install --upgrade pip
-          pip install -e "python[all]"
-          pip install transformers==4.45.2
-          pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/ --force-reinstall
+          bash scripts/ci_install_dependency.sh
 
       - name: Benchmark Single Latency
         timeout-minutes: 10
@@ -165,10 +147,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          pip install --upgrade pip
-          pip install -e "python[all]"
-          pip install transformers==4.45.2
-          pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/ --force-reinstall
+          bash scripts/ci_install_dependency.sh
 
       - name: Benchmark Offline Throughput (w/o RadixAttention)
         timeout-minutes: 10
@@ -197,10 +176,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          pip install --upgrade pip
-          pip install -e "python[all]"
-          pip install transformers==4.45.2
-          pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/ --force-reinstall
+          bash scripts/ci_install_dependency.sh
 
       - name: Benchmark Offline Throughput (TP=2)
         timeout-minutes: 10
@@ -229,10 +205,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          pip install --upgrade pip
-          pip install -e "python[all]"
-          pip install transformers==4.45.2
-          pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/ --force-reinstall
+          bash scripts/ci_install_dependency.sh
 
           git clone https://github.com/merrymercy/human-eval.git
           cd human-eval
@@ -253,10 +226,7 @@ jobs:
 
       - name: Install dependencies
         run: |
-          pip install --upgrade pip
-          pip install -e "python[all]"
-          pip install transformers==4.45.2
-          pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/ --force-reinstall
+          bash scripts/ci_install_dependency.sh
 
           git clone https://github.com/merrymercy/human-eval.git
           cd human-eval

.github/workflows/release-github.yml

@@ -1,25 +0,0 @@
-name: Release GitHub
-on:
-  workflow_dispatch:
-jobs:
-  publish:
-    if: github.repository == 'sgl-project/sglang'
-    runs-on: ubuntu-latest
-    environment: 'prod'
-    steps:
-      - name: Checkout repository
-        uses: actions/checkout@v3
-      
-      - name: Get version
-        id: get_version
-        run: |
-          version=$(cat python/sglang/version.py | cut -d'"' -f2)
-          echo "TAG=v$version" >> $GITHUB_OUTPUT
-
-      - name: Release
-        uses: softprops/action-gh-release@v1
-        env:
-          GITHUB_TOKEN: ${{ secrets.REPO_TOKEN }}
-        with:
-          name: Release ${{ steps.get_version.outputs.TAG }}
-          tag_name: ${{ steps.get_version.outputs.TAG }}

.github/workflows/release-pypi.yml

@@ -13,12 +13,14 @@ jobs:
     runs-on: ubuntu-latest
     environment: 'prod'
     steps:
-      - name: Set up python3.8
+      - name: Set up Python
         uses: actions/setup-python@v4
         with:
-          python-version: '3.8'
+          python-version: '3.9'
+
       - name: Checkout repository
         uses: actions/checkout@v3
+
       - name: Upload to pypi
         run: |
           cd python

docs/Makefile

@@ -0,0 +1,23 @@
+# Minimal makefile for Sphinx documentation
+#
+
+# You can set these variables from the command line, and also
+# from the environment for the first two.
+SPHINXOPTS    ?=
+SPHINXBUILD   ?= sphinx-build
+SOURCEDIR     = .
+BUILDDIR      = _build
+
+# Put it first so that "make" without argument is like "make help".
+help:
+	@$(SPHINXBUILD) -M help "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O)
+
+.PHONY: help Makefile
+
+# Catch-all target: route all unknown targets to Sphinx using the new
+# "make mode" option.  $(O) is meant as a shortcut for $(SPHINXOPTS).
+%: Makefile
+	@$(SPHINXBUILD) -M $@ "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O)
+
+clean:
+	rm -rf $(BUILDDIR)/*

docs/README.md

@@ -0,0 +1,33 @@
+# SGLang Documentation
+
+## Build the documentation website
+
+### Dependency
+```
+pip install -r requirements.txt
+```
+
+### Build
+```
+make html
+```
+
+### Clean
+To remove all generated files:
+```
+make clean
+```
+
+### Serve (preview)
+Run an HTTP server and visit http://localhost:8000 in your browser.
+```
+python3 -m http.server --d _build/html
+```
+
+### Deploy
+Clone [sgl-project.github.io](https://github.com/sgl-project/sgl-project.github.io) and make sure you have write access.
+
+```bash
+export DOC_SITE_PATH=../../sgl-project.github.io   # update this with your path
+python3 deploy.py
+```

docs/backend.md

@@ -20,7 +20,7 @@ curl http://localhost:30000/generate \
   }'
 ```
 
-Learn more about the argument specification, streaming, and multi-modal support [here](https://sglang.readthedocs.io/en/latest/sampling_params.html).
+Learn more about the argument specification, streaming, and multi-modal support [here](https://sgl-project.github.io/sampling_params.html).
 
 ### OpenAI Compatible API
 In addition, the server supports OpenAI-compatible APIs.
@@ -74,7 +74,7 @@ python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct
 ```
 python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --mem-fraction-static 0.7
 ```
-- See [hyperparameter tuning](https://sglang.readthedocs.io/en/latest/hyperparameter_tuning.html) on tuning hyperparameters for better performance.
+- See [hyperparameter tuning](https://sgl-project.github.io/hyperparameter_tuning.html) on tuning hyperparameters for better performance.
 - If you see out-of-memory errors during prefill for long prompts, try to set a smaller chunked prefill size.
 ```
 python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct --chunked-prefill-size 4096
@@ -83,7 +83,7 @@ python -m sglang.launch_server --model-path meta-llama/Meta-Llama-3-8B-Instruct
 - To enable torchao quantization, add `--torchao-config int4wo-128`. It supports various quantization strategies.
 - To enable fp8 weight quantization, add `--quantization fp8` on a fp16 checkpoint or directly load a fp8 checkpoint without specifying any arguments.
 - To enable fp8 kv cache quantization, add `--kv-cache-dtype fp8_e5m2`.
-- If the model does not have a chat template in the Hugging Face tokenizer, you can specify a [custom chat template](https://sglang.readthedocs.io/en/latest/custom_chat_template.html).
+- If the model does not have a chat template in the Hugging Face tokenizer, you can specify a [custom chat template](https://sgl-project.github.io/custom_chat_template.html).
 - To run tensor parallelism on multiple nodes, add `--nnodes 2`. If you have two nodes with two GPUs on each node and want to run TP=4, let `sgl-dev-0` be the hostname of the first node and `50000` be an available port, you can use the following commands. If you meet deadlock, please try to add `--disable-cuda-graph`
 ```
 # Node 0
@@ -158,7 +158,7 @@ You can view the full example [here](https://github.com/sgl-project/sglang/tree/
 - gte-Qwen2
   - `python -m sglang.launch_server --model-path Alibaba-NLP/gte-Qwen2-7B-instruct --is-embedding`
 
-Instructions for supporting a new model are [here](https://sglang.readthedocs.io/en/latest/model_support.html).
+Instructions for supporting a new model are [here](https://sgl-project.github.io/model_support.html).
 
 #### Use Models From ModelScope
 <details>

docs/conf.py

@@ -3,7 +3,7 @@ import sys
 
 sys.path.insert(0, os.path.abspath("../.."))
 
-version_file = "../../python/sglang/version.py"
+version_file = "../python/sglang/version.py"
 with open(version_file, "r") as f:
     exec(compile(f.read(), version_file, "exec"))
 __version__ = locals()["__version__"]

docs/contributor_guide.md

@@ -11,4 +11,4 @@ pre-commit run --all-files
 ```
 
 ## Add Unit Tests
-Add unit tests under [sglang/test](../../test). You can learn how to add and run tests from the README.md in that folder.
+Add unit tests under [sglang/test](https://github.com/sgl-project/sglang/tree/main/test). You can learn how to add and run tests from the README.md in that folder.

docs/custom_chat_template.md

@@ -1,6 +1,6 @@
 # Custom Chat Template in SGLang Runtime
 
-**NOTE**: There are two chat template systems in SGLang project. This document is about setting a custom chat template for the OpenAI-compatible API server (defined at [conversation.py](../../python/sglang/srt/conversation.py)). It is NOT related to the chat template used in the SGLang language frontend (defined at [chat_template.py](../../python/sglang/lang/chat_template.py)).
+**NOTE**: There are two chat template systems in SGLang project. This document is about setting a custom chat template for the OpenAI-compatible API server (defined at [conversation.py](https://github.com/sgl-project/sglang/blob/main/python/sglang/srt/conversation.py)). It is NOT related to the chat template used in the SGLang language frontend (defined at [chat_template.py](https://github.com/sgl-project/sglang/blob/main/python/sglang/lang/chat_template.py)).
 
 By default, the server uses the chat template specified in the model tokenizer from Hugging Face.
 It should just work for most official models such as Llama-2/Llama-3.

docs/deploy.py

@@ -0,0 +1,22 @@
+#!/usr/bin/python3
+
+import os
+from datetime import datetime
+
+
+def run_cmd(cmd):
+    print(cmd)
+    os.system(cmd)
+
+
+run_cmd("cd $DOC_SITE_PATH; git pull")
+
+# (Optional) Remove old files
+# run_cmd("rm -rf $ALPA_SITE_PATH/*")
+
+run_cmd("cp -r _build/html/* $DOC_SITE_PATH")
+
+cmd_message = f"Update {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}"
+run_cmd(
+    f"cd $DOC_SITE_PATH; git add .; git commit -m '{cmd_message}'; git push origin main"
+)

docs/en/Makefile

@@ -1,12 +0,0 @@
-SPHINXOPTS    =
-SPHINXBUILD   = sphinx-build
-SOURCEDIR     = .
-BUILDDIR      = _build
-
-help:
-	@$(SPHINXBUILD) -M help "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O)
-
-.PHONY: help Makefile
-
-%: Makefile
-	@$(SPHINXBUILD) -M $@ "$(SOURCEDIR)" "$(BUILDDIR)" $(SPHINXOPTS) $(O)

docs/index.rst

@@ -10,24 +10,28 @@ The core features include:
 - **Extensive Model Support**: Supports a wide range of generative models (Llama 3, Gemma 2, Mistral, QWen, DeepSeek, LLaVA, etc.) and embedding models (e5-mistral), with easy extensibility for integrating new models.
 - **Active Community**: SGLang is open-source and backed by an active community with industry adoption.
 
+
 .. toctree::
    :maxdepth: 1
    :caption: Getting Started
 
    install.md
-   send_request.ipynb
+
 
 .. toctree::
    :maxdepth: 1
    :caption: Backend Tutorial
+
    backend.md
 
 
 .. toctree::
    :maxdepth: 1
    :caption: Frontend Tutorial
+
    frontend.md
 
+
 .. toctree::
    :maxdepth: 1
    :caption: References
@@ -39,4 +43,3 @@ The core features include:
    choices_methods.md
    benchmark_and_profiling.md
    troubleshooting.md
-   embedding_model.ipynb

docs/install.md

@@ -48,9 +48,9 @@ docker run --gpus all \
 <summary>More</summary>
 
 > This method is recommended if you plan to serve it as a service.
-> A better approach is to use the [k8s-sglang-service.yaml](./docker/k8s-sglang-service.yaml).
+> A better approach is to use the [k8s-sglang-service.yaml](https://github.com/sgl-project/sglang/blob/main/docker/k8s-sglang-service.yaml).
 
-1. Copy the [compose.yml](./docker/compose.yaml) to your local machine
+1. Copy the [compose.yml](https://github.com/sgl-project/sglang/blob/main/docker/compose.yaml) to your local machine
 2. Execute the command `docker compose up -d` in your terminal.
 </details>
 

docs/requirements.txt

@@ -1,12 +1,17 @@
+ipykernel
+ipywidgets
+jupyter_client
 markdown>=3.4.0
+matplotlib
 myst-parser
+nbconvert
+nbsphinx
+pandoc
+pillow
+pydantic
 sphinx
 sphinx-book-theme
 sphinx-copybutton
 sphinx-tabs
 sphinxcontrib-mermaid
-pillow
-pydantic
-urllib3<2.0.0
-nbsphinx
-pandoc
+urllib3<2.0.0

docs/sampling_params.md

@@ -194,7 +194,7 @@ Since we compute penalty algorithms through CUDA, the logic stores relevant para
 
 You can run your own benchmark with desired parameters on your own hardware to make sure it's not OOMing before using.
 
-Tuning `--mem-fraction-static` and/or `--max-running-requests` will help. See [here](hyperparameter_tuning.md#minor-tune---max-prefill-tokens---mem-fraction-static---max-running-requests) for more information.
+Tuning `--mem-fraction-static` and/or `--max-running-requests` will help.
 
 ### Benchmarks
 

docs/serve.sh

@@ -0,0 +1 @@
+python3 -m http.server --d _build/html

docs/troubleshooting.md

@@ -5,9 +5,9 @@ This page lists some common errors and tips for fixing them.
 ## CUDA error: an illegal memory access was encountered
 This error may be due to kernel errors or out-of-memory issues.
 - If it is a kernel error, it is not easy to fix.
-- If it is out-of-memory, sometimes it will report this error instead of "Out-of-memory." In this case, try setting a smaller value for `--mem-fraction-static`. The default value of `--mem-fraction-static` is around 0.8 - 0.9. https://github.com/sgl-project/sglang/blob/1edd4e07d6ad52f4f63e7f6beaa5987c1e1cf621/python/sglang/srt/server_args.py#L92-L102
+- If it is out-of-memory, sometimes it will report this error instead of "Out-of-memory." In this case, try setting a smaller value for `--mem-fraction-static`. The default value of `--mem-fraction-static` is around 0.8 - 0.9.
 
 ## The server hangs
 If the server hangs, try disabling some optimizations when launching the server.
 - Add `--disable-cuda-graph`.
-- Add `--disable-flashinfer-sampling`.
+- Add `--sampling-backend pytorch`.

scripts/ci_install_dependency.sh

@@ -0,0 +1,4 @@
+pip install --upgrade pip
+pip install -e "python[all]"
+pip install transformers==4.45.2
+pip install flashinfer -i https://flashinfer.ai/whl/cu121/torch2.4/ --force-reinstall