v0.1.7

Updates in this new release v0.1.7 are:

• New source code example for the paper Grouped Pointwise Convolutions Reduce Parameters in Convolutional Neural Networks has been added.

  

• Updated readme file and source code comments.
• Plenty of minor source code improvements and fixes.
• New function cai.models.CreatePartialModelWithSoftMax: creates a partial model up to the layer name defined in pOutputLayerName and then adds a dense layer with softmax. This method is used for creating a smaller model from an existing model.
• New function cai.models.CreatePartialModelFromChannel: creates a partial model up to the layer name defined in pOutputLayerName and then copies the channel at index pChannelIdx. This method is used for creating a model that computes the output of a given channel from an existing bigger model.
• New function cai.datasets.extract_subset_every: deterministic function to move a subset of the dataset (files) to another folder.
• New function cai.datasets.clone_sub_folders: creates new images by flipping horizontally and vertically the existing images.

New Function cai.datasets.save_tfds_in_format

The new function cai.datasets.save_tfds_in_format saves a TensorFlow dataset as image files. Classes become folders for easy of use by standard Keras API.

kEffNet 1.0

Basic code for the paper "Grouped Pointwise Convolutions Significantly Reduces Parameters in EfficientNet".

Heatmaps and Deep Dream

Added method to display heatmaps: cai.models.calculate_heat_map_from_dense_and_avgpool.

Added method to calculate patterns from inner layers via Gradient Ascent / Deep Dream: cai.gradientascent.run_gradient_ascent_octaves.

Requires and supports tensorflow 2.2.0

v0.0.9

Porting code to TF2.2.0.

Plenty of coding

v0.0.8

Update simple_image_classification_with_any_dataset.ipynb

Bug fixing for mnist dataset.

v0.0.2

Update README.md