Aditya Kane

Aditya Kane

Computer Vision and Natural Language Processing

My open source contributions

This page summarizes my open source contributions. All of them can be found on GitHub.

KerasCV

KerasCV is a repository of modular building blocks (layers, metrics, losses, data-augmentation) that applied computer vision engineers can leverage to quickly assemble production-grade, state-of-the-art training and inference pipelines.

My contributions to KerasCV include the following Pull Requests. Some of the important ones include RandomResizedCrop (a.k.a. Inception crop, Stochastic Depth and porting over models from keras.applications to keras_cv.models.

Keras

Keras is Deep learning API built on top of TensorFlow. Keras offers consistent & simple APIs, it minimizes the number of user actions required for common use cases, and it provides clear & actionable error messages. It also has extensive documentation and developer guides.

My contributions to Keras include the following Pull Requests. Most notably, I have added RegNets to tf.keras.applications. Check out the PR here.

TensorFlow

TensorFlow is an end-to-end open source platform for machine learning.

My contributions to TensorFlow include the following Pull Requests. To summarize, my contributions to this project are as follows:

  1. Added “RegNetY” (see below) to TFHub.
  2. Removed few bugs reported by users.
  3. Improved documentation.

I also contributed by resolving issues on their GitHub page as well as StackOverflow.

My contributions to Google’s ML Ecosystem are recorded in this document.

Google Summer of Code (GSoC)

Google Summer of Code is a global, online program focused on bringing new contributors into open source software development. GSoC Contributors work with an open source organization on a 12+ week programming project under the guidance of mentors. I was a GSoC student contributor at TensorFlow for 2021 and 2022 programs.

GSoC 2021 at TensorFlow

My project included implementing and training RegNetY models from the paper Designing network design spaces by Radosavovic et al. on ImageNet-1k dataset. The final report is available here.

GSoC 2022 at TensorFlow

My project included incorporating several model blocks and augmentation layers to KerasCV. The final project report is available here.

Other projects

I have contributed to small projects like pytorch-retinanet (a PyTorch implementation of the RetinaNet), SmallPebble (a see-through autodiff library written in Python), awesomo (a collection of interesting projects).