Downloads · 30 days
22
13% of all-time downloads
researchpaper-classifier
vighneshkhot/researchpaper-classifier
researchpaper-classifier is a machine learning model from vighneshkhot. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
"A deep learning NLP model that automatically classifies research paper abstracts into their correct academic category, built using DistilBERT and PyTorch, trained on diverse and varied data." ---
All-time downloads
175
Public
Parameters
67M
268 MB on disk
Likes
1
Public
Repo makeup
Click a slice to open those files.
.safetensors268 MB · 100%
At a glance
- Model type
- distilbert
- Access
- Public
- Created
- Apr 10, 2026
- Updated
- Oct 14, 2026
- SHA
- 42fb0961
Base models
Model card
From the Hugging Face model README
📄 Research Paper Classifier
"A deep learning NLP model that automatically classifies research paper abstracts into their correct academic category, built using DistilBERT and PyTorch, trained on diverse and varied data."
🚀 Project Overview
Every day, thousands of research papers are published across multiple academic fields. This project builds an automated classifier that reads a paper abstract and instantly predicts its category — such as Computer Science, Mathematics, Physics, Statistics, and more.
This project demonstrates the application of Natural Language Processing (NLP) and deep learning for academic document classification.
📊 Results
| Metric | Score |
|---|---|
| Test Accuracy | 82% |
| Weighted F1 Score | 0.8133 |
| Training Samples | ~3000,000 |
| Categories | 20 |
| Training Environment | VS Code (Local GPU) |
🏆 Best Performing Categories
| Category | F1 Score |
|---|---|
| Astrophysics | 0.94 |
| Computer Science | 0.89 |
| Mathematics | 0.89 |
| Quantum Physics | 0.80 |
📂 Dataset
- Source: Cornell University arXiv dataset
- Total Available: ~3 million papers
- Used for Training: ~3000,000 samples
- Input: Research Paper Details
- Output: 20 academic categories
🤖 Model Details
- Base Model: DistilBERT (base-uncased)
- Task: Sequence Classification
- Max Token Length: 128
- Epochs: 3
- Learning Rate: 2e-5
🛠️ Tech Stack
- Python
- PyTorch
- Hugging Face Transformers
- Scikit-learn
- VS Code (Local GPU)
- Type
- distilbert
- Created
- Apr 10, 2026
- Updated
- Oct 14, 2026