Doof245/Monet_Style_Transfer_Pix2Pix_GAN

Monet Style Transfer with Pix2Pix GAN

Transform your photos into Monet-style paintings using a trained Pix2Pix GAN.

Model Architecture

• —Generator: U-Net with skip connections (7 encoder + 7 decoder layers)
• —Discriminator: PatchGAN (classifies 30x30 patches)
• —Loss Functions: Adversarial Loss + L1 Pixel Loss

Training Details

• —Dataset: Monet paintings and landscape photos from Kaggle
• —Epochs: 30-100
• —Framework: PyTorch
• —Image Size: 256x256

Usage

1. 1.Upload any photo (landscapes work best)
2. 2.Click "Generate Monet Style"
3. 3.Download your artistic result

Examples

Works best with:

• —Landscape photos
• —Nature scenes
• —Outdoor photography
• —Images with clear subjects

Model Performance

The model learns to apply Monet's impressionist style while preserving the content and structure of the input photo.