CVNets

When it fits

Use CVNets when your work involves training or evaluating computer-vision networks in a Python and PyTorch workflow and you want repository examples spanning several tasks and model families.

Who it is for

• Researchers developing or comparing computer-vision models.
• Engineers who need a Python toolkit for training classification, detection or segmentation networks.
• Developers exploring mobile vision models or CLIP workflows.

Documented uses

• Train image-classification models, including listed CNN and transformer families.
• Train object detectors such as SSD or Mask R-CNN.
• Train semantic-segmentation models such as DeepLabv3 or PSPNet.
• Explore augmentation and distillation techniques in model-training workflows.
• Use the provided guidance to explore conversion of PyTorch models to CoreML.