A modular workspace for developing, testing, and deploying chatbots using various models, datasets, and templates.
chonkie/
– Dataset tools and utilities.models.dev/
– Model zoo, demos, or model code.llama_factory/
– Model loading and inference. (Example repo you may add)easy_dataset/
– Dataset utilities. (Example repo you may add)autotrain-advanced/
– Advanced training scripts. (Example repo you may add)templates/
demo-chatbot/
– Example or starter chatbot template.- (Add more templates in separate subfolders for isolation and easy deployment.)
main.py
– Entry point for integration or orchestration.
- Place each new chatbot template in its own folder under
templates/
for isolation and easy deployment. - Add models, datasets, or other utilities as needed.
- Document each template in its own README inside the subfolder.
To add a new template:
- Create a new folder under
templates/
, e.g.templates/my-new-bot/
. - Add your code, config (e.g.
main.py
,config.yaml
,requirements.txt
). - Include a
README.md
explaining usage and deployment.
You can add more code repositories to your chatbot-factory
directory to extend functionality (e.g. model loading, training, dataset management).
To add a new repo:
From inside your chatbot-factory
directory, run:
git clone https://github.com/owner/repo-name.git
This will create a new folder called repo-name/
in your project.
If you want to track remote updates separately:
git submodule add https://github.com/owner/repo-name.git repo-name
This will also create a new folder with the repo’s code.
You can also manually copy code folders into chatbot-factory/
if you prefer.
After adding a repo:
- Update the Directory Structure section in this README.
- Add any usage notes or integration details as needed.
chatbot-factory/
├── chonkie/
├── llama_factory/
├── easy_dataset/
├── autotrain-advanced/
├── templates/
│ ├── demo-chatbot/
│ └── orderbot/
├── library/
│ ├── models.dev/
│ └── RAG_Techniques/
├── agents
│ ├── mastra/
│ ├── AutoGroq/
├── main.py
└── README.md
python main.py
Then explore, add, or run your bots!
MIT (or specify your license)