This repository contains a variety of utils I use in day-to-day work. Between the deep learning frameworks, I use torch so some utils will require this dependency.
Hope you find something useful or interesting inside!
Utilities for using the computer/laptop camera in your computer vision applications.
Loaders and converters for images, pointclouds, meshes and more. Also tools for changing the dimension order of image tensors and image concatenation.
Generic data structure solutions for very specific use-cases.
A flexible implementation of logging utilities for your generic Python needs. Construct your own logging pipeline by encapsulating the leaf loggers with various decorators. The global LOG is a singleton logger instance that can be accessed anywhere by just importing it (no need for passing loggers as a method argument).
Loss tracking and model checkpointing to reduce code clutter in optimization loops.
Experiment tracking and hyperparameter searches for easier project management and development. Also other utils such as GIF generator.
Various utils for plotting various results using matplotlib. Also, plotting contexts for cleaner and more reliable plotting code.
General tools such as a timer and memory management.
Torch specific tools, mainly regarding the experiment reproducibility (seeding of random generators).
A bunch of convolution kernels an various other tools that are in the domain of computer vision.