trickyy00045/Qwen47

Fine-Tuned Student Profile Assistant

This repository contains a QLoRA/LoRA fine-tuning project based on Qwen/Qwen2.5-0.5B-Instruct.

Profile covered

• Name: Stephen Dhanush F
• College: Dayananda Sagar University
• Program: MSc Data Science

Repository files

• app.py — Gradio chatbot for Hugging Face Spaces.
• train.py — reproducible QLoRA fine-tuning script.
• training_data.json — training examples.
• training_notebook.ipynb — original notebook.
• requirements.txt — Python dependencies.
• README.md — project documentation.

Training

Use a CUDA-enabled GPU:

pip install -r requirements.txt
python train.py

The LoRA adapter and tokenizer are saved in:

./stephen_dhanush_profile_model/

Hugging Face model repository

After training, upload the contents of stephen_dhanush_profile_model/ to your Hugging Face model repository. The adapter files are generated by training and are not included in this source package.

Hugging Face Space

For a Gradio Space, keep app.py, requirements.txt, and the trained adapter directory available to the Space. If the adapter is stored in a separate Hugging Face model repository, download/snapshot it into the Space at ./stephen_dhanush_profile_model before starting the app.

Base model

Qwen/Qwen2.5-0.5B-Instruct

This project uses LoRA/PEFT rather than saving a complete copy of the base model.