sentiment-imdb-distilbert

Model Description

Fine-tuned distilbert-base-uncased for binary sentiment classification on the IMDB movie review dataset.

Training Date: 2025-11-18 14:27:43

Intended Use

This model classifies text into positive or negative sentiment. It was trained on movie reviews but may generalize to other domains.

Performance

Metric	Score
Accuracy	0.9304
F1 Score	0.9306
Precision	0.9283
Recall	0.9330

Training Hyperparameters

• Model: distilbert-base-uncased
• Epochs: 4
• Batch Size: 32
• Learning Rate: 2e-05
• Max Length: 512
• Warmup Ratio: 0.1
• Weight Decay: 0.01
• Training Samples: Full dataset (25,000)

Usage

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

model_name = "None/sentiment-imdb-distilbert" if args.hf_username else "local model"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name)

def predict(text):
    inputs = tokenizer(text, return_tensors="pt", truncation=True, max_length=512)
    with torch.no_grad():
        outputs = model(**inputs)
    probs = torch.nn.functional.softmax(outputs.logits, dim=-1)
    return "Positive" if probs[0][1] > 0.5 else "Negative"

print(predict("This movie was amazing!"))  # Positive

Limitations

• Trained primarily on movie reviews; performance may vary on other text types
• May reflect biases present in the IMDB dataset
• English language only

Citation

@misc{sentiment-imdb-distilbert,
  author = {Your Name},
  title = {Sentiment Analysis with distilbert-base-uncased},
  year = {2025},
  publisher = {HuggingFace},
  howpublished = {\url{https://huggingface.co/None/sentiment-imdb-distilbert}}
}