PyTorch Retinanet
A PyTorch implementation of RetinaNet for object detection as described in Focal Loss for Object Detection.
Hi! 👋 I'm Ayushman
Hi! I am Ayushman, a 3rd Year Undergrad student currently pursuing a Bachelor of Technology in
Computer Science Engineering at the Central Institute of Technology , Kokrajhar.
My research interests are in Deep Learning, Computer Vision and Reinforcement Learning.
I am originally from the small city of Guwahati, Assam in India which hosts the smallest river
island in the world.
I enjoy coding especially in Python. My github account hosts all of my code and documentation,
which includes all the notebooks from the blog posts as well as any projects
There are quite a few things I am passionate about in life, and coding is just one facet of my
life. I enjoy reading novels in my free-time especially the ones with thriller and mystery
elements in it.
Skills
Programming Languages
Python Libraries
Personal Projects
These are a few of the notable projects that has been done by me. All the projects/code are uploaded to my Github.
Another Deep Learning Blog
This is my personal-blog created using Fastpages and GitHub Actions.
Retinanet Pet Detector
- A RetinaNet model to detect faces of pets (cats and dogs) in images.
- This model was trained on the Oxford-IIT Pets Dataset.
- Used StreamLit to create a simple web interface and it was deployed here.
OneCycleLearningRate for TensorFlow2
- Implemented few aspects of the paper Super-Convergence: Very Fast Training of Neural Networks Using Large Learning Rates .
- Created a tf.keras callback which implements the OneCycle Policy.
- Also implemented a LearningRateFinder, which was also described in the above mentioned paper.