unipercept 2.1.1

Tools and libraries for unified multi-task vision research.

Unipercept

Public repository of tools and libraries powering cutting-edge computer vision and machine learning research at the Mobile Perception Systems (MPS) lab at Eindhoven University of Technology (TU/e).

Logging

Importing unipercept does not attach any console handler by default, so no output will appear on stdout/stderr until you explicitly enable it.

Enabling console output from your application

Call unipercept.log.configure() once at your entry point:

import unipercept.log

unipercept.log.configure()  # attach console handler to the root unipercept logger

Enabling console output via environment variable

Set UNIPERCEPT_LOG_AUTO_CONFIGURE=1 before the first import to have configure() called automatically:

UNIPERCEPT_LOG_AUTO_CONFIGURE=1 python my_script.py

Controlling log levels

The default level is INFO. Override it globally with UP_LOGS_LEVEL:

UP_LOGS_LEVEL=debug python my_script.py

Per-module overrides use UP_LOGS (comma-separated):

# Explicit level
UP_LOGS="unipercept.nn:debug,unipercept.ops:warning" python my_script.py

# Relative adjustment (+/- around the base level)
UP_LOGS="+unipercept.nn,-unipercept.ops" python my_script.py