fine-tune-inference-gui

Cheeky little Tkinter GUI for fine-tuning LLMs - suited for a dev environment.

Must have an NVIDIA GPU for CUDA compatibility.

This tool:

1. Assumes you have enough storage for your fine-tuned LLM as it saves it locally.
2. Can pull from HuggingFace, so you need to have a token for the specified model; this can be configured through the HuggingFace CLI
3. Must be configured directly in the code to change model (unless you want to use GPT-2)

Features

• Fine-tuning
• Inference

GUI

Dependencies

• PyTorch
• Transformers
• Datasets

pip install transformers torch datasets

Run using:

python fine_tune_gui.py