/llm-runpod

Primary LanguagePython

Superflows RunPod setup

This repo is for using the Superflows fine-tuned LLM on https://runpod.io.

You want to download the model weights from hugging face and put all model-0000X-of-0000Y.safetensors files in the /zephyr directory.

The adapter_model.safetensors file should live in the /superflows directory.

These files aren't included in this repo because they are too large.

Deployment

You need to:

  1. make a RunPod account
  2. build the docker container using the Dockerfile in this repo
  3. push the docker container to a container registry
  4. deploy the container to RunPod

Reach out to henry@superflows.ai if you have any questions or need help with this process.