README.md

---
license: apache-2.0
language:
- en
metrics:
- FID
- IS
pipeline_tag: unconditional-image-generation
tags:
- generative-adversarial-network
- pytorch
- dcgan
- deep-learning
---

# Model Card for DCGAN (PyTorch)

## Model Details

### Model Description

This is a **Deep Convolutional Generative Adversarial Network (DCGAN)** implemented in **PyTorch**.  
It is trained to generate synthetic images that resemble the target dataset distribution.  

- **Developed by:** Abhishek C.  
- **Funded by [optional]:** Self Funded
- **Shared by:** None
- **Model type:** Generative Adversarial Network (DCGAN)  
- **Language(s):** N/A (Image generation)  
- **License:** Apache-2.0  
- **Finetuned from model [optional]:** Not applicable (trained from scratch)  
---

## Uses

### Direct Use
- Generating synthetic images from random noise vectors (`z ~ N(0,1)`).
- Data augmentation for research and experimentation.  
- Educational purposes to study GAN training and generative modeling.  

### Downstream Use
- Fine-tuning the discriminator or generator on domain-specific datasets.  
- Using the pretrained generator as an initialization for conditional GANs.  

### Out-of-Scope Use
- Medical or safety-critical applications without validation.  
- Misuse for generating harmful or misleading content.  

---

## Bias, Risks, and Limitations
- Generated images may contain artifacts if training is insufficient.  
- Quality depends heavily on dataset diversity and size.  
- Model may amplify dataset biases.  

### Recommendations
- Always evaluate generated images before downstream use.  
- Do not use in decision-critical tasks.  
- Use larger datasets for stable performance.  

---

## How to Get Started with the Model

```python
import torch
from torch import nn

# Load pretrained generator (example structure)
class Generator(nn.Module):
    def __init__(self, nz=100, ngf=64, nc=3):
        super().__init__()
        self.main = nn.Sequential(
            nn.ConvTranspose2d(nz, ngf*8, 4, 1, 0, bias=False),
            nn.BatchNorm2d(ngf*8),
            nn.ReLU(True),
            nn.ConvTranspose2d(ngf*8, ngf*4, 4, 2, 1, bias=False),
            nn.BatchNorm2d(ngf*4),
            nn.ReLU(True),
            nn.ConvTranspose2d(ngf*4, ngf*2, 4, 2, 1, bias=False),
            nn.BatchNorm2d(ngf*2),
            nn.ReLU(True),
            nn.ConvTranspose2d(ngf*2, nc, 4, 2, 1, bias=False),
            nn.Tanh()
        )

    def forward(self, input):
        return self.main(input)

# Example usage
netG = Generator()
noise = torch.randn(16, 100, 1, 1)
fake_images = netG(noise)