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.

Python client for Kili Technology labeling tool. Kili Technology is an image, text and voice data annotation tool designed to help companies deploy machine learning applications faster. In a few minutes you can start annotating your data thanks to a catalogue of intuitive and configurable interfaces. You can easily accelerate the labeling process by connecting one of your models to pre annotate the data. The work of the annotators is 2 to 5 times faster. Kili Technology facilitates collaboration between technical teams and the business, but also with outsourced annotation companies. Data governance is managed, and production quality control is facilitated. Kili Technology meets the needs of small teams as well as those of large companies with massive stakes.

Semi-Automatic OCR Annotation Tool

Statistical test to detect higher-order correlations between count variables.

A package to convert your Jupyter Notebook

Open Chemometrics Toolkit for Analysis and Visualization of Vibrational Spectroscopy data

Quickest change detection algorithms for online streaming data in python. The onlineChange python package is designed to quickest detect any change in distributions for online streaming data, supporting any user-specified distributions. It supports sequential Monte Carlo to perform Bayesian change point analysis, where the ‘parameters before or after the change can be unknown’. It also supports likelihood ratio based test to determine the stopping time (i.e. when change occurs).

Keras extension for Openml python API

MXNet extension for Openml python API

Distributed Training Package Based on PaddlePaddle. Fleet Package is a high level distributed training package. A user can easily define deep learning models with fleet built-in package or generate a new model with API. Online Training is supported in Fleet package.

Interact with PLTs

A package to eliminate multicollinearity

A high performance implementation of Numenta’s HTM algorithms