Scanning all new published packages on PyPI I know that the quality is often quite bad. I try to filter out the worst ones and list here the ones which might be worth a look, being followed or inspire you in some way.

Collect and clean data using Excel spreadsheets.

Machine learning for systems

A collection of layers and utils for TensorFlow (Keras) 2.+

package to ease and speed up machine learning model explainability

A package for doing hyper-spectral image augmentation for deep learning. Other image augmentation libraries often only accept RGB or grayscale images (with options for RGBA images usually too). However, in a scientific context, we often have images with more than 3/4 channels (so-called hyper-spectral imaging) and it seemed like a good idea to have an augmentation library which deals with all channels in parallel. Another problem is that some image augmentation frameworks can accept more than 3/4 channels but convert the data to unsigned 8-bit integers (see PyTorch’s (https://…/transforms.html )) which is damaging for scientific data where we care about the actual numbers of the data. This can lead to the images losing some of the features and relative contrasts of features which is important for our science. I had a look at several image augmentation packages but none of them seemed to satisfy both of these criteria so here we are.

Dataset loaders for pytorch

Wrap your multiple torch.Tensor`s into single TensorStruct and use it like you are using torch.Tensor.

Deep Learning toolbox for WSI (digital histopatology) analysis

Tools for geographical data processing

High quality, fast, modular reference implementation of SSD in PyTorch