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.

• jax-loss
A jax library of common machine learning loss functions

• nvision
Nipa Cloud Platform: Nvision Python SDK

• openbiolink
A framework for evaluating link prediction models on heterogeneous biomedical graph data. OpenBioLink is a resource and evaluation framework for evaluating link prediction models on heterogeneous biomedical graph data. It contains benchmark datasets as well as the underlying scrips to create them and to evaluate a costume model on them.

• photosynthesis-metrics
Measures and metrics for image2image tasks. PyTorch.

• reformer-pytorch
Reformer, the Efficient Transformer, Pytorch

• rhodb
Rho AI Database Helpers

• robotpy-halsim-lowfi
WPILib command framework

• StatArbTools
A set of tools useful in exploring statistical arbitrage. StatArbTools is a Python library primarily for determining if a pair of time series are cointegrated. It also includes tools for generating an array of log returns from a price array, looking for a linear relationship, and creating a potentially stationary distribution.

• stylegan2-pytorch
StyleGan2 in Pytorch

• voicenotes2org
Transcribe voice recordings using the Google Cloud Speech-To-Text API, and export the results to Emacs org-mode headings.

• zero-divisor-graph
a library for working with zero divisor graphs of commutative semigroups

• arfs-gen
Library to generate toy data for machine learning experiments in the context of all-relevant feature selection.