FunGPT (TinyGPT) — Hugging Face Model Card
FunGPT is a tiny decoder-only Transformer (GPT-style) trained for playful, snarky, and creative text generation. It uses a SentencePiece tokenizer and a small 6x384 architecture for quick, lightweight inference.
Quick Start
clone this repo: git clone https://huggingface.co/nagolinc/fungpt
run chat: python chat.py
reproduction
the model can be reproduced using train2.py
- Downloads last month
- 90
Install from pip and serve model
# Install vLLM from pip: pip install vllm# Start the vLLM server: vllm serve "nagolinc/fungpt"# Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "nagolinc/fungpt", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'