https://github.com/tddschn/modelz-starcoder
https://github.com/tddschn/modelz-starcoder
Last synced: 9 days ago
JSON representation
- Host: GitHub
- URL: https://github.com/tddschn/modelz-starcoder
- Owner: tddschn
- Created: 2023-06-14T08:38:32.000Z (almost 3 years ago)
- Default Branch: main
- Last Pushed: 2023-06-15T09:00:21.000Z (almost 3 years ago)
- Last Synced: 2026-04-16T21:06:09.094Z (about 2 months ago)
- Language: Python
- Size: 18.6 KB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Modelz MusicGen
[StarCoder](https://huggingface.co/bigcode/starcoder) is a language model (LM) trained on source code and natural language text. Its training data incorporates more that 80 different programming languages as well as text extracted from GitHub issues and commits and from notebooks.
This repository contains the code (Dockerfile, [`build.envd`](https://envd.tensorchord.ai/guide/getting-started.html)) for deploying MusicGen on [Modelz](https://docs.modelz.ai/).
## Getting Started
This is a template for creating a [Gradio](https://gradio.app/) app on [Modelz](https://modelz.ai/).
Building an Gradio app could be straightforward. You will need to provide three key components:
- A `main.py` file: This file contains the code for making predictions.
- A `requirements.txt` file: This file lists all the dependencies required for the server code to run.
- A `Dockerfile` or a simpler [`build.envd`](https://envd.tensorchord.ai/guide/getting-started.html): This file contains instructions for building a Docker image that encapsulates the server code and its dependencies.
## Build
In the `Dockerfile`, you need to define the instructions for building a Docker image that encapsulates the server code and its dependencies.
In most cases, you could use the template in the repository.
```bash
docker build -t docker.io/USER/IMAGE .
docker push docker.io/USER/IMAGE
```
On the other hand, a [`build.envd`](https://envd.tensorchord.ai/guide/getting-started.html) is a simplified alternative to a Dockerfile. It provides python-based interfaces that contains configuration settings for building a image.
It is easier to use than a Dockerfile as it involves specifying only the dependencies of your machine learning model, not the instructions for CUDA, conda, and other system-level dependencies.
```bash
envd build --output type=image,name=docker.io/USER/IMAGE,push=true
```
## Deploy
Please refer to the [Modelz documentation](https://docs.modelz.ai/gettingstarted/deploy) for more details.