group4-sentiment-model

Sentiment classifier (positive / negative) fine-tuned from distilbert-base-uncased.

Group members: Orishaba Jenipher and Nalweyiso Racheal Dataset: stanfordnlp/imdb (movie reviews)

Results on the test set

Metric	Score
Accuracy	0.90
F1	0.90

Example predictions

• "Terrible experience, I want my money back." → NEGATIVE (0.86)
• "The food was amazing!" → POSITIVE (0.99)
• "It was okay, nothing special." → NEGATIVE (0.73)

Limitations

Trained on English movie reviews only; may struggle with sarcasm, Luganda text, and mixed-language reviews.