- Creating the "ImagLoaderTranslateTorch" class for loading images in batches based on file paths:
- When initializing the class with the "train" mode, images are selected and then augmented during loading.
- The "get_class_distribution" function is used to display the number of images belonging to each class.
- The "num_class" function is used to display the number of classes in the sample.
- The "imshow" function is used to display the image on the screen.
- Creating variables for the "ImagLoaderTranslateTorch" class.
- Testing random images.
- Building a network based on ResNet using Fine-Tuning.
- Creating functions for training the network.
- Training and testing the network (Training on training data, visualizing results, visualizing images with predictions, further training on validation data, submitting).
- Recognizing a single real image (functions for recognizing a single real loaded image).
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|