rahulporuri / deeplabcut-core

Headless DeepLabCut (no GUI support)

DeepLabCut-core

JUNE 2021: THIS CODE IS NOW DEPRECIATED! DeepLabCut now supports 2.2 for standard, multi-animal, and DeepLabCut-Live! See main repo for details!

UPDATE JAN 2021: We will be using this space as the tensorflow 2 test-bed, but then all DeepLabCut will be within the main package. The headless version will be pip install deeplabcut, while the full GUI supported version is pip install deeplabcut[gui]. This means deeplabcutcore will be depreciated once TF2 is merged into the main repo

Currently up to date with DeepLabCut v2.1.8.1. AND uses tensorflow 2.X

Core functionalities of DeepLabCut, excluding all GUI functions.

Please be aware, you can create projects, etc. with the full deeplabcut package. Here, you will need to create the training set and train, evaluate, etc. not inter-mixing with using the deeplabcut package (it currently supports tensorflow 1.x). We recommend looking at this google colab notebook to help you, and this blog post about our transitition to tensorflow 2.

Install from GitHub: pip install git+https://github.com/DeepLabCut/DeepLabCut-core

PyPi: pip install deeplabcutcore

Documentation is located at DeepLabCut's main GitHub page.