CenterNet is a framework for object detection with deep convolutional neural networks. You can use the code to train and evaluate a network for object detection on the MS-COCO dataset. It achieves state-of-the-art performance (an AP of 47.0%) on one of the most challenging dataset: MS-COCO. Our code is written in Python, based on CornerNet. More detailed descriptions of our approach and code will be made available soon. If you encounter any problems in using our code, please contact Kaiwen Duan: firstname.lastname@example.org.
This repository contains the search and evaluation code for our work Progressive DARTS. It requires only 0.3 GPU-days (7 hours on a single P100 card) to finish a search progress on CIFAR10 and CIFAR100 datasets, much faster than DARTS, and achieves higher classification accuracy on both CIFAR and ImageNet datasets (mobole setting).
PC-DARTS is a memory-efficient differentiable architecture method based on DARTS. It mainly focuses on reducing the large memory cost of the super-net in one-shot NAS method, which means that it can also be combined with other one-shot NAS method e.g. ENAS. Different from previous methods that sampling operations, PC-DARTS samples channels of the constructed super-net. Interestingly, though we introduced randomness during the search process, the performance of the searched architecture is better and more stable than DARTS! For a detailed description of technical details and experimental results, please refer to our paper: Partial Channel Connections for Memory-Efficient Differentiable Architecture Search Yuhui Xu, Lingxi Xie, Xiaopeng Zhang, Xin Chen, Guo-Jun Qi, Qi Tian and Hongkai Xiong. This code is based on the implementation of DARTS.