Implementation:SqueezeAILab ETS Sglang ETS Installation

Knowledge Sources	ETS sglang-ets
Domains	Infrastructure, Setup
Last Updated	2026-02-14 02:00 GMT

Overview

Concrete tool for installing the ETS experiment runtime dependencies via shell commands.

Description

This implementation covers the shell commands needed to clone and install the sglang-ets custom fork and its required companion library outlines. The sglang-ets fork extends the standard SGLang serving framework with tree search state forking, PRM score backends, and batched multi-threaded inference needed by the ETS algorithm.

Usage

Run these commands once when setting up a new environment for ETS experiments. Must be executed before launching any SGLang model servers.

Code Reference

Source Location

• Repository: ETS
• File: README.md
• Lines: 19-28

Signature

# Clone and install sglang-ets fork
git clone https://github.com/chooper1/sglang-ets.git
cd sglang-ets/python
pip install .

# Install pinned outlines version
pip install outlines==0.0.44

Import

# After installation, sglang is available as:
from sglang import function, gen, RuntimeEndpoint, system, user, assistant

I/O Contract

Inputs

Name	Type	Required	Description
Git URL	str	Yes	Repository URL: https://github.com/chooper1/sglang-ets.git
outlines version	str	Yes	Pinned version: 0.0.44

Outputs

Name	Type	Description
sglang package	Python package	Installed sglang module with ETS extensions (state forking, score backends, run_batch)
outlines package	Python package	Installed outlines library for structured generation

Usage Examples

Full Installation

# 1. Clone the custom sglang fork
git clone https://github.com/chooper1/sglang-ets.git

# 2. Install sglang-ets
cd sglang-ets/python
pip install .

# 3. Install outlines (pinned version)
pip install outlines==0.0.44

# 4. Verify installation
python3 -c "from sglang import function, gen, RuntimeEndpoint; print('sglang-ets installed successfully')"

Related Pages

Implements Principle

• Principle:SqueezeAILab_ETS_Dependency_Installation

Requires Environment

• Environment:SqueezeAILab_ETS_Multi_GPU_Sglang_Runtime