YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

Frox AI Morph Nano

A 207M parameter language model trained from scratch on WikiText-103. Built with the Frox Morph 1.1 architecture.

Quick Start

pip install -r requirements.txt
python app.py

The app downloads the model from HuggingFace Hub automatically and starts a Gradio chat interface.

HF Token

Set your HF token to download the private model:

export HF_TOKEN=hf_your_token_here
python app.py

Deploy to HF Space

To create a HuggingFace Space:

  1. Go to https://huggingface.co/new-space
  2. Name: frox-nano-1-1
  3. SDK: Gradio
  4. Space Hardware: CPU (free)
  5. Upload all files from this repo
  6. Set HF_TOKEN as a Space secret
  7. Space will auto-start and download the model

Training

To train from scratch:

python pretrain_nano.py

Files

  • app.py - Gradio chat app (entry point)
  • src/ - Full Frox Morph 1.1 source code (74 modules)
  • model_best/ - Trained model weights + config + tokenizer
  • requirements.txt - Python dependencies
  • pretrain_nano.py - Training script
Downloads last month
219
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support