https://github.com/sferadev/fine-tuner

https://github.com/sferadev/fine-tuner

Last synced: 27 days ago
JSON representation

• Host: GitHub
• URL: https://github.com/sferadev/fine-tuner
• Owner: SferaDev
• Created: 2023-10-09T17:27:34.000Z (about 1 year ago)
• Default Branch: main
• Last Pushed: 2023-10-09T19:28:07.000Z (about 1 year ago)
• Last Synced: 2024-10-09T10:53:57.860Z (about 1 month ago)
• Language: TypeScript
• Size: 17.6 KB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# fine-tuner

A small framework to build, train and deploy fine-tuned OpenAI models.

## Create a chat session

You can create a chat session by running:

```bash

$ fine-tuner chat

```

The session will be stored in `sessions/timestamp.json` and will look like this:

```json

{

  "messages": []

}

```

You can manually edit the sessions to add messages or change the system messages.

## Create a fine-tuned model

The stored sessions can be used to create a fine-tuned model by running:

```bash

$ fine-tuner deploy

```

If the model was deployed previously, it will be updated instead.

## Refresh sessions

If you want to refresh the sessions, you can run:

```bash

$ fine-tuner refresh

```

All function calls will be executed again and the assisant will be trained with the new data.

This is useful if you want to change a function or update the system messages.

## Configuration

The configuration is stored in `config.json` and looks like this:

```json

{

  "model": "gpt-3.5-turbo",

  "id": "fine-tune-model-id",

  "system_messages": []

}

```