Vision backbone networks serves as the feature extractor of input image data.
This repository is dedicated to the exploration of compelling and influential deep learning-based vision foundational backbones, neck and head concepts. The primary objective is to derive inspiration from research papers, translating these ideas into code to gain a comprehensive understanding of the proposed concepts. I believe that by reading and implementing the vision backbone networks proposed by the research community, we too can formulate our own ideas and do better engineering work instead of treating the outcome of research papers as plug and play.
My core focus will be studying the inspiration of the paper, the core contribution made and the implementation in code.
Most of the source code are inspired from open-mmlab and awesome-backbones
@repo{2020mmclassification,
title={OpenMMLab's Image Classification Toolbox and Benchmark},
author={MMClassification Contributors},
howpublished = {\url{https://github.com/open-mmlab/mmclassification}},
year={2020}
}