https://github.com/abdbarho/transformers-stack
A full stack solution for deploying a transformers model from HuggingFace
https://github.com/abdbarho/transformers-stack
docker docker-compose huggingface nextjs transformers
Last synced: about 1 year ago
JSON representation
A full stack solution for deploying a transformers model from HuggingFace
- Host: GitHub
- URL: https://github.com/abdbarho/transformers-stack
- Owner: AbdBarho
- Created: 2022-08-21T08:51:50.000Z (almost 4 years ago)
- Default Branch: master
- Last Pushed: 2022-11-11T06:24:10.000Z (over 3 years ago)
- Last Synced: 2025-04-30T20:08:07.820Z (about 1 year ago)
- Topics: docker, docker-compose, huggingface, nextjs, transformers
- Language: JavaScript
- Homepage:
- Size: 79.1 KB
- Stars: 3
- Watchers: 2
- Forks: 3
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Transformers Stack
A full stack solution for deploying a language model from hugging face with a simple UI for prompting the model and tracking the results.
This project uses `docker compose` for orchestrating two containers, `model` which contains the model with a simple http interface, and `ui` which is a next.js application that provides access to the model in a UI.
## Setup
update the variable `MODEL_KEY` in `docker-compose.yml` to the name of the model you want, for example `bigscience/bloom-560m` or `EleutherAI/gpt-neo-1.3B`.
And then run `docker compose up --build`! thats all you need! The UI will be available on [http://localhost:9999](http://localhost:9999) (you can change the port also in `docker-compose.yml` )
### Run on CPU
create a file called `docker-compose.override.yml` and put the following in it:
```yml
services:
model:
deploy: {}
```