This project is based on li-plus/chatglm.cpp to package a default Xorbits/chatglm2-6B-GGML chatglm2-ggml-q4_1.bin model. It can be run in a Docker container, allowing you to experience chatglm2 directly through your browser, using only CPU, out of the box.
Run the following command using Docker
docker run -p 7860:7860 -d zhangp365/chatglm-webui:v0.1
Access it in your browser at http://localhost:7860/
You can replace it with other models supported by li-plus/chatglm.cpp. When starting Docker, use the -v parameter and specify the model name in the -e environment variable. Place the model in the corresponding external directory as follows:``
docker run -p 7860:7860 -e CLI_ARGS="-m models/your_model_name" -v /your/path/models:/app/models -d zhangp365/chatglm-webui:v0.1
Link: https://hub.docker.com/repository/docker/zhangp365/chatglm-webui/general Size: packaging a 3.9GB chatglm2-ggml-q4_1.bin model, after compression in Docker, the image size is only 3.7GB, making it very convenient to download and experience.