HangulAI_Model - GitLite

KO | EN

gitlite — search

#large-language-models #mcp #chatbot #proxy #generative-ai #android #llm #sql #golang #redis

HangulAI_Model

(★ 11)

AI model for recognizing Korean Hangul characters using a CNN in PyTorch. Incorporates data augmentation, batch normalization, dropout, and AdamW optimizer. Supports scalable training, early stopping, and fast inference for OCR and language-learning applications.

File Explorer

ZIP Download

- - 4008.png
  - 5515.png
- - 3401.png
CNN.py
dataset.py
LICENSE
model.pth
README.md
requirements.txt
tend.png
testing.py

# Use via CDN

jsDelivr

jsDelivr serves any public GitHub repository as a CDN with zero setup. Pick a version and a file to get a ready-to-paste link and snippet.

Version

File

Link

Example

Command Glossary

Commands referenced in this DOCs, explained below.

🔍

`pip install`

View Details ▼

Install Python packages.

pip install {{package1 package2 ...}}

Install one or more packages:

pip install {{package1 package2 ...}} {{[-U|--upgrade]}}

Upgrade all specified packages to the latest version, installing any that are not already present:

pip install {{package}}=={{version}}

Install a specific version of a package:

🔍

`python`

View Details ▼

Python language interpreter.

python

Start a REPL (interactive shell):

python {{path/to/file.py}}

Execute a specific Python file:

python -i {{path/to/file.py}}

Execute a specific Python file and start a REPL:

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

A modern CNN-based solution for recognizing Korean Hangul characters with state-of-the-art training techniques.
<p align="center">
  <img src="https://raw.githubusercontent.com/LokiChan69/HangulAI_Model/main/tend.png" alt="Neural Network Architecture" width="90%">
  <br>
  <em>Training and validation accuracy/loss curves for the Hangul character recognition model</em>
</p>

## ✨ Key Features

| Feature                       | Description                                   |
| ----------------------------- | --------------------------------------------- |
| **Advanced CNN Architecture** | ResNet-inspired with BatchNorm and Dropout    |
| **Optimized Training**        | AdamW optimizer + ReduceLROnPlateau scheduler |
| **Smart Augmentation**        | Rotation, flipping, color jittering           |
| **Early Stopping**            | Prevents overfitting automatically            |
| **Model Checkpointing**       | Always saves best performing model            |

## 🛠️ Technical Specifications

### Model Architecture
```python
CNN(
  (features): Sequential(
    Conv2d → BatchNorm2d → ReLU → MaxPool2d → Dropout
    [... repeated blocks ...]
  )
  (classifier): Linear → LogSoftmax
)
```

### Training Parameters
```yaml
image_size: 64x64
batch_size: 64
epochs: 50
optimizer: AdamW (lr=0.001)
loss: CrossEntropyLoss
scheduler: ReduceLROnPlateau 
early_stopping: patience=5
```

## 📂 Project Structure

```
.
├── 📁 dataset/           # Training data (organized by class)
│   ├── 가/               # Example class folder
│   │   ├── img1.jpg
│   │   └── img2.jpg
│   └── 나/
├── 📄 CNN.py             # Main training script
├── 📄 testing.py         # For testing the model
├── 📄 model.pth          # Saved model weights
└── 📄 requirements.txt   # Dependencies
```

## 🚀 Quick Start

1. **Setup Environment**
```bash
pip install -r requirements.txt
```

2. **Organize Dataset**
```
dataset/
  ├── class_1/
  │   ├── image1.jpg
  │   └── image2.jpg
  └── class_2/
```

3. **Start Training**
```bash
python train.py --gpu  # Add --gpu for GPU acceleration
```

## 🎯 Performance Metrics

![Training Progress](https://via.placeholder.com/400x200?text=Training+Metrics+Chart) *(example visualization)*

Expected performance:
- Training Accuracy: ~98%
- Validation Accuracy: ~95%
- Inference Speed: 15ms/image (on GPU)

## 📜 License

MIT License - Free for academic and commercial use

---

> "Perfect for Korean OCR applications, educational tools, and language learning apps" - *AI Research Team*