Here I tried to make a set of templates for different tasks in image recognition and computer vision using the TorchVision package.
- Binary Classification (CNN from scratch, Sigmoid Activation, BinaryCrossEntropyLoss)
- MultiClass Classification (CNN from scratch, ReLu, CrossEntropy)
- MultiClass Classification (CNN from pre-trained models)
- Object Detection (find object label, position, mask)
Read instruction in each folder's README file for each task's acceptable input data format.