metadata

language: en
tags:
  - sentiment-analysis
  - modernbert
  - imdb
datasets:
  - imdb
metrics:
  - accuracy
  - f1

ModernBERT IMDb Sentiment Analysis Model

Model Description

Fine-tuned ModernBERT model for sentiment analysis on IMDb movie reviews. Achieves 95.75% accuracy on the test set.

Usage

from transformers import AutoModelForSequenceClassification, AutoTokenizer

model = AutoModelForSequenceClassification.from_pretrained("voxmenthe/modernbert-imdb-sentiment")
tokenizer = AutoTokenizer.from_pretrained("answerdotai/ModernBERT-base")

# Input processing
inputs = tokenizer("This movie was fantastic!", return_tensors="pt")
outputs = model(**inputs)

# Get the predicted class
predicted_class_id = outputs.logits.argmax().item()

# Convert class ID to label
predicted_label = model.config.id2label[predicted_class_id]
print(f"Predicted label: {predicted_label}")

Model Card

Model Details

Model Name: ModernBERT IMDb Sentiment Analysis
Base Model: answerdotai/ModernBERT-base
Task: Sentiment Analysis
Dataset: IMDb Movie Reviews
Training Epochs: 5

Model Performance

Test Accuracy: 95.75%
Test F1 Score: 95.75%

Model Architecture

Base Model: answerdotai/ModernBERT-base
Task-Specific Head: ClassifierHead (from classifiers.py)
Number of Labels: 2 (Positive, Negative)

Model Inference

Input Format: Text (single review)
Output Format: Predicted sentiment label (Positive or Negative)

Model Version

Version: 1.0
Date: 2025-05-07

Model License

License: MIT License

Model Contact

Contact: [email protected]

Model Citation

Citation: voxmenthe/modernbert-imdb-sentiment