Qwen is an official Alibaba-backed Python repository hosting Qwen base LLMs and chat models, with extensive documentation and integration examples. It notes that the repository is no longer actively maintained.
Collecting history — the radar snapshots this repo daily. The trend line appears after 3 days of data (1 so far).
What it is
The repository provides Qwen base language models (Qwen-1.8B, Qwen-7B, Qwen-14B, Qwen-72B) and Qwen-Chat variants. It offers links to model cards, a technical report, and a FAQ for issues. It states that Qwen2 exists and that this repo is no longer actively maintained.
How it works
The README describes model cards and deployment options, including quantization details (GPTQ, KV cache), inference performance statistics, and tutorials for finetuning (full-parameter, LoRA, Q-LoRA). It includes deployment instructions with vLLM and FastChat, and demonstrates demos via WebUI and CLI. It also covers DashScope API service and tool/agent usage.
Getting started
- Requirements: python 3.8+, PyTorch 1.12+ (2.0+ recommended), transformers 4.32+, CUDA 11.4+.
- Quickstart includes: "pip install -r requirements.txt" and optional flash-attention installation steps.
- Commands shown:
pip install -r requirements.txt
git clone https://github.com/Dao-AILab/flash-attention
cd flash-attention && pip install .
- Example usage for inference with 🤗 Transformers, requiring model names such as "Qwen/Qwen-7B-Chat" and "Qwen/Qwen-14B-Chat" and code snippet showing model.chat usage.
Recent releases
- RELEASES (latest 0): - none
Traction
- The repo list shows a high star count (21546) but the README itself provides no separate traction metrics beyond the repository overview.
Behind the repo
- The README mentions that this repo is no longer actively maintained and points to Qwen2 as a separate repo for activity.
Caveats
- License: Apache-2.0
- Created: 2023-08-03; last_push: 2026-03-05
- Open issues: 47
- The README notes substantial codebase differences and that the project is not actively maintained.