|
|
--- |
|
|
license: mit |
|
|
language: |
|
|
- en |
|
|
metrics: |
|
|
- accuracy |
|
|
base_model: |
|
|
- google-bert/bert-base-uncased |
|
|
pipeline_tag: text-classification |
|
|
tags: |
|
|
- sentiment-analysis |
|
|
- bert |
|
|
- imdb |
|
|
- fine-tuned |
|
|
- text-classification |
|
|
--- |
|
|
|
|
|
# 📊 Sentiment Analysis with Fine-Tuned BERT (IMDB) |
|
|
|
|
|
This repository contains a fine-tuned BERT model for binary sentiment classification using the IMDB movie reviews dataset. The model classifies reviews as **positive** or **negative**, and is built using Hugging Face Transformers and PyTorch. |
|
|
|
|
|
## 🚀 Model Performance |
|
|
|
|
|
| Metric | Value | |
|
|
|------------------|-------------| |
|
|
| Accuracy | 89.4% | |
|
|
| Validation Loss | 0.375 | |
|
|
| Epochs Trained | 3 | |
|
|
| Inference Speed | ~434 samples/sec | |
|
|
|
|
|
## 🧠 Model Details |
|
|
|
|
|
- **Base Model**: `bert-base-uncased` |
|
|
- **Dataset**: IMDB (binary sentiment) |
|
|
- **Framework**: Hugging Face Transformers |
|
|
- **Fine-Tuning Setup**: |
|
|
- Learning rate: 2e-5 |
|
|
- Batch size: 32 |
|
|
- Mixed-precision: ✅ (`fp16`) |
|
|
- Early stopping: ❌ (trained for full 3 epochs) |
|
|
|
|
|
## 🛠️ How to Use |
|
|
|
|
|
```python |
|
|
from transformers import pipeline |
|
|
|
|
|
classifier = pipeline("text-classification", model="Harsha901/tinybert-imdb-sentiment-analysis-model") |
|
|
classifier("This movie was absolutely amazing!") |
|
|
|
