C++ implementation of Qwen-LM for real-time chatting on your MacBook.

• 2023/12/05 qwen was merged to llama.cpp and supports gguf format.

Highlights:

• Pure C++ implementation based on ggml, working in the same way as llama.cpp.
• Pure C++ tiktoken implementation.
• Streaming generation with typewriter effect.
• Python binding.

Support Matrix:

• Hardwares: x86/arm CPU, NVIDIA GPU
• Platforms: Linux, MacOS
• Models: Qwen-LM

Preparation

Clone the qwen.cpp repository into your local machine:

git clone --recursive https://github.com/QwenLM/qwen.cpp && cd qwen.cpp

If you forgot the --recursive flag when cloning the repository, run the following command in the qwen.cpp folder:

git submodule update --init --recursive

Download the qwen.tiktoken file from Hugging Face or modelscope.

Quantize Model

Use convert.py to transform Qwen-LM into quantized GGML format. For example, to convert the fp16 original model to q4_0 (quantized int4) GGML model, run:

python3 qwen_cpp/convert.py -i Qwen/Qwen-7B-Chat -t q4_0 -o qwen7b-ggml.bin

The original model (-i <model_name_or_path>) can be a HuggingFace model name or a local path to your pre-downloaded model. Currently supported models are:

• Qwen-7B: Qwen/Qwen-7B-Chat
• Qwen-14B: Qwen/Qwen-14B-Chat

You are free to try any of the below quantization types by specifying -t <type>:

• q4_0: 4-bit integer quantization with fp16 scales.
• q4_1: 4-bit integer quantization with fp16 scales and minimum values.
• q5_0: 5-bit integer quantization with fp16 scales.
• q5_1: 5-bit integer quantization with fp16 scales and minimum values.
• q8_0: 8-bit integer quantization with fp16 scales.
• f16: half precision floating point weights without quantization.
• f32: single precision floating point weights without quantization.

Build & Run

Compile the project using CMake:

cmake -B build
cmake --build build -j --config Release

Now you may chat with the quantized Qwen-7B-Chat model by running:

./build/bin/main -m qwen7b-ggml.bin --tiktoken Qwen-7B-Chat/qwen.tiktoken -p 你好
# 你好！很高兴为你提供帮助。

To run the model in interactive mode, add the -i flag. For example:

./build/bin/main -m qwen7b-ggml.bin --tiktoken Qwen-7B-Chat/qwen.tiktoken -i

In interactive mode, your chat history will serve as the context for the next-round conversation.

Run ./build/bin/main -h to explore more options!

OpenBLAS

OpenBLAS provides acceleration on CPU. Add the CMake flag -DGGML_OPENBLAS=ON to enable it.

cmake -B build -DGGML_OPENBLAS=ON && cmake --build build -j

cuBLAS

cuBLAS uses NVIDIA GPU to accelerate BLAS. Add the CMake flag -DGGML_CUBLAS=ON to enable it.

cmake -B build -DGGML_CUBLAS=ON && cmake --build build -j

Metal

MPS (Metal Performance Shaders) allows computation to run on powerful Apple Silicon GPU. Add the CMake flag -DGGML_METAL=ON to enable it.

cmake -B build -DGGML_METAL=ON && cmake --build build -j

The Python binding provides high-level chat and stream_chat interface similar to the original Hugging Face Qwen-7B.

Installation

Install from PyPI (recommended): will trigger compilation on your platform.

pip install -U qwen-cpp

You may also install from source.

# install from the latest source hosted on GitHub
pip install git+https://github.com/QwenLM/qwen.cpp.git@master
# or install from your local source after git cloning the repo
pip install .

We provide pure C++ tiktoken implementation. After installation, the usage is the same as openai tiktoken:

import tiktoken_cpp as tiktoken
enc = tiktoken.get_encoding("cl100k_base")
assert enc.decode(enc.encode("hello world")) == "hello world"

Benchmark

The speed of tiktoken.cpp is on par with openai tiktoken:

cd tests
RAYON_NUM_THREADS=1 python benchmark.py

Unit Test

To perform unit tests, add this CMake flag -DQWEN_ENABLE_TESTING=ON to enable testing. Recompile and run the unit test (including benchmark).

mkdir -p build && cd build
cmake .. -DQWEN_ENABLE_TESTING=ON && make -j
./bin/qwen_test

Lint

To format the code, run make lint inside the build folder. You should have clang-format, black and isort pre-installed.

• This project is greatly inspired by llama.cpp, chatglm.cpp, ggml, tiktoken, tokenizer, cpp-base64, re2 and unordered_dense.