How to Build a Docker Image

To build a Docker image, use the following command:

docker build -t mygirlgpt .

How to Run MyGirlGPT in RunPod

Assuming you already have a stable diffusion service with API enabled.

To run in RunPod, follow these steps:

Add a new template and fill in the required information, including the template name, container image (use synthintel2/mygirlgpt:latest or your own built image), volume disk (if needed), and exposed HTTP ports (use 7860 for text-generation-webui and 8888 for Jupyter).
Add the required environment variables, including LOAD_MODEL (default value is TehVenom/Pygmalion-Vicuna-1.1-7b), OPENAI_API_KEY, SD_ADDRESS, and BOT_TOKEN.
Optionally, you can add additional environment variables, such as OPENAI_API_BASE and JUPYTER_PASSWORD.

Here's an example of how to fill in the fields when adding a new template:

Start the server by selecting the template and clicking "Continue". Wait for the pod to start up and become ready.

Here's an example of how to start a pod once the template is ready:

Once it's ready, you can start chatting with Cherry.