https://github.com/zyuanlim/genai-nd-project-1
https://github.com/zyuanlim/genai-nd-project-1
Last synced: 3 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/zyuanlim/genai-nd-project-1
- Owner: zyuanlim
- Created: 2024-08-09T10:01:50.000Z (10 months ago)
- Default Branch: main
- Last Pushed: 2024-08-11T08:31:18.000Z (10 months ago)
- Last Synced: 2025-01-16T15:43:29.023Z (5 months ago)
- Language: Jupyter Notebook
- Size: 35.2 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# GenAI Nanodegree Project 1: Apply Lightweight Fine-Tuning to a Foundation Model
Code walkthrough: [Notebook](<./Lightweight%20Fine-Tuning%20Foundation%20Model%20(deberta-v3).ipynb>)
## Task Overview
Classify whether the news is `real` (0) or `fake` (1). The dataset is from [HuggingFace](https://huggingface.co/datasets/mohammadjavadpirhadi/fake-news-detection-dataset-english).
## Model Overview
The model is the [deberta-v3](https://huggingface.co/microsoft/deberta-v3-base) from Microsoft.
## Fine-Tuning
The model is fine-tuned using QLoRA.
## Model Performance
- The original (non-fine-tuned) model has an accuracy of 53.6% on the test set.
- The fine-tuned model has an accuracy of 100% on the test set.