|
|
--- |
|
|
title: Answer Evaluation App |
|
|
emoji: 🧠 |
|
|
colorFrom: blue |
|
|
colorTo: green |
|
|
sdk: docker |
|
|
app_port: 7860 |
|
|
--- |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
# Answer Evaluation App ✍️ |
|
|
|
|
|
# 📝 Handwritten Answer Evaluation App using OCR + XLNet |
|
|
|
|
|
An end-to-end machine learning application that extracts handwritten answers from images and evaluates them using a custom XLNet model trained on semantic similarity. |
|
|
|
|
|
--- |
|
|
|
|
|
## 🚀 Live Demo |
|
|
|
|
|
👉 [Try it on Hugging Face Spaces](https://yeswanthvarma-answer-evaluation-app.hf.space) |
|
|
|
|
|
--- |
|
|
|
|
|
## 📌 Project Overview |
|
|
|
|
|
This app takes images of handwritten answers and performs: |
|
|
1. **OCR** to extract text from question, student answer, and reference answer. |
|
|
2. **Similarity scoring** using a custom-trained XLNet model. |
|
|
3. **Bonus logic** to adjust the final score based on thresholds. |
|
|
4. **User interface** to upload images and view the evaluated score. |
|
|
|
|
|
--- |
|
|
|
|
|
## 🧠 Core Technologies |
|
|
|
|
|
- **FastAPI**: Web framework |
|
|
- **EasyOCR**: For extracting handwritten text |
|
|
- **Hugging Face Transformers**: XLNet model |
|
|
- **Custom Training**: Trained on Q-A-R triplets |
|
|
- **Docker**: For containerized deployment |
|
|
- **Hugging Face Spaces**: Live hosted app |
|
|
|
|
|
--- |
|
|
|
|
|
## 📦 Folder Structure |
|
|
|
|
|
``` |
|
|
answer-evaluation-app/ |
|
|
├── app.py # FastAPI application |
|
|
├── requirements.txt # Dependencies |
|
|
├── Dockerfile # For Hugging Face deployment |
|
|
├── utils/ |
|
|
│ ├── image_processor.py # EasyOCR + preprocessing |
|
|
│ └── xlnet_model.py # Model load and prediction |
|
|
├── templates/ |
|
|
│ └── index.html # Frontend HTML |
|
|
├── static/ |
|
|
│ ├── css/style.css # UI styling |
|
|
│ ├── js/main.js # JS for client interaction |
|
|
│ └── uploads/ # Uploaded image storage |
|
|
``` |
|
|
|
|
|
--- |
|
|
|
|
|
## 🔍 Model Details |
|
|
|
|
|
- **Base Model**: `xlnet-base-cased` (Hugging Face) |
|
|
- **Custom Trained On**: Question, student answer, reference answer, and human-evaluated scores |
|
|
- **Loss**: MSELoss |
|
|
- **Output**: Score from 0 to 100 |
|
|
--- |
|
|
|
|
|
## ✍️ Sample Use Case |
|
|
|
|
|
- Upload 3 images: |
|
|
- Question image |
|
|
- Student handwritten answer |
|
|
- Reference answer |
|
|
- App will: |
|
|
- Extract text |
|
|
- Score similarity using model |
|
|
- Apply bonus logic |
|
|
- Display final score and extracted text |
|
|
|
|
|
--- |
|
|
|
|
|
## 🧑💻 Author |
|
|
|
|
|
**Yeswanth Varma Gottumukkala** |
|
|
- Email: [email protected] |
|
|
|
|
|
--- |
|
|
|
|
|
## 📜 License |
|
|
|
|
|
This project is for educational and research purposes. |
|
|
Model and app are freely available to explore on Hugging Face Spaces. |
