File size: 1,790 Bytes
941a7fe |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 |
---
license: mit
language:
- en
tags:
- gan
- mnist
- 7gen
- pytorch
library_name: torch
model_type: image-generator
---
# 7Gen - Advanced MNIST Digit Generation System
**State-of-the-art Conditional GAN for MNIST digit synthesis with self-attention mechanisms.**
---
## 🚀 Features
- 🎯 **Conditional Generation**: Generate specific digits (0–9) on demand.
- 🖼️ **High Quality Output**: Sharp and realistic handwritten digit samples.
- ⚡ **Fast Inference**: Real-time generation on GPU.
- 🔌 **Easy Integration**: Minimal setup, PyTorch-native implementation.
- 🚀 **GPU Acceleration**: Full CUDA support.
---
## 🔍 Model Details
- **Architecture**: Conditional GAN with self-attention
- **Parameters**: 2.5M
- **Input**: 100-dimensional noise vector + class label
- **Output**: 28x28 grayscale images
- **Training Data**: MNIST dataset (60,000 images)
- **Training Time**: ~2 hours on NVIDIA RTX 3050 Ti
---
## 🧪 Performance Metrics
| Metric | Score |
|------------------|-------|
| **FID Score** | 12.3 |
| **Inception Score** | 8.7 |
- **Training Epochs**: 100
- **Batch Size**: 64
---
## ⚙️ Training Configuration
```yaml
model:
latent_dim: 100
num_classes: 10
generator_layers: [256, 512, 1024]
discriminator_layers: [512, 256]
training:
batch_size: 64
learning_rate: 0.0002
epochs: 100
optimizer: Adam
beta1: 0.5
beta2: 0.999
|