SLEAP v0.2.1

This is the first non-develop release of v0.2.x.

SLEAP v0.2 has major changes (from v0.1.x) to training/inference code and uses TensorFlow 2.1. It can read datasets from v0.1, although it cannot currently use models or training profiles.

SLEAP v0.2.1a12

Automated release of 0.2.1a12 on 2020-04-08

SLEAP v0.2.1a11

Automated release of 0.2.1a11 on 2020-04-06

SLEAP v0.2.1a10

Automated release of 0.2.1a10 on 2020-04-06

SLEAP v0.2.1a9

Automated release of 0.2.1a9 on 2020-04-01

SLEAP v0.2.1a8

Automated release of 0.2.1a8 on 2020-03-31 (manual rebuild)

SLEAP v0.2.1a7

Automated release of 0.2.1a7 on 2020-03-31

SLEAP v0.2.1a6

• Bug fix for adding missing points when point isn't present in "template" instance
• Option to predict on all user labeled frames or random frames in current video.

SLEAP v0.2.1a5

Conda build changes:

• require tensorflow-gpu=2.1 from conda
• remove most of the host requirements for building package
• add tf to run requirements for built package

SLEAP v0.2.1a4

Install tensorflow 2.1 from conda.