Image Segmentation Model Trainer
Upload Training Data
Upload a ZIP file containing image pairs for training. Each pair should consist of a color image and its corresponding segmentation mask (e.g., 00001_color.png and 00001_mask.png).
Training Progress
Controls how much the model changes with each training step. Higher values (0.01-0.03): Faster learning but may be unstable. Lower values (0.001-0.005): More stable but slower training.
Number of image pairs processed in each training batch. Larger batches (70-100): Better generalization but uses more memory. Smaller batches (10-30): Less memory usage but may have more variance.