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.

anomaly-detection-ts
anomaly detection sdk

cloudframe
A light set of supporting modules to assist the data science workflow based on Cloudframe’s proprietary data science enablers

deepctr-torch
Easy-to-use,Modular and Extendible package of deep learning based CTR(Click Through Rate) prediction models with tensorflow.

jupyter-images
package for generating HTML code to put images in Jupyter notebooks, especially from google drive

jupyter-notebookparams
Pass URL parameters to a Jupyter notebook. Takes query parameters from a url to update a parameter cell of a jupyter notebook.

ml-finance-tools
Machine Learning Tools useful for Financial Time Series. This package is designed to provide useful classes for machine learning in finance. There are many unique issues in financial machine learning and in time series data in general.

nas
A library for Network Architecture Search (NAS)

nestedtensor
NestedTensors for PyTorch

sklearn-lmer
Scikit-learn estimator wrappers for pymer4 wrapped LME4 mixed effects models

video-facenet
Face detection/embeddings/clustering for video files using Google’s FaceNet deep neural network & TensorFlow.

xlnet-tensorflow
XLNet for TensorFlow. This is a fork of the original [XLNet repository](https://…/xlnet ) that adds package configuration so that it can be easily installed and used. The purpose is to remove the need of cloning the repository and modifying it locally which can be quite dirty for common tasks (e.g. training a new classifier). A lot of code can be shared (e.g. `modeling.py`, `classifier_utils.py`) and this fork is exactly for that.

abc_algorithm
Artificial Bee Colony Algorithm for educational purposes

archives
a new way to do python code documentation

astra-forge
A library for training PyTorch models.

deipy
Data Analysis Engine

Advertisements