sentiment-tutorial

Fine-tuned distilbert-base-uncased for binary sentiment classification.

Intended Use

Classify English text as positive or negative.

Training Procedure

• Base model: distilbert-base-uncased
• Epochs: 2
• Learning rate: 2e-5
• Batch size: 32
• Max length: 128

Evaluation Results

Accuracy: 0.870

Precision: 0.879

Recall: 0.858

F1: 0.868

Limitations

• Binary classification only
• English only
• Movie reviews domain

Usage

from transformers import pipeline

classifier = pipeline( "sentiment-analysis", model="ayesha9f/sentiment-tutorial" )

classifier("This was a great experience!")