WebGPU-based Image-to-Image Translation (Beta)
Input Images
Drag & drop folder or click to select
0 filesOutput Images (Ground Truth)
Drag & drop folder or click to select
0 filesU-Net Training
Train a ResNet34-based U-Net model on your mosaic dataset
Dataset
Training samples:
0
Validation samples:
0
Hyperparameters
Model Architecture
Training Progress
Epoch 0 / 0
Train Loss
-
Val Loss
-
Learning Rate
-
Time/Epoch
-
Training Preview (waiting...)
Input
Prediction
Target
Inference
Apply trained model to new images with tiled processing and weighted blending
No model loaded
Drop Image Here
or click to select
Left: Input | Right:
Result
Upload an image to automatically run inference