MUXConv: Information Multiplexing in Convolutional Neural Networks

Benchmark Model Rank Results
architecture-search-on-cifar-10-imageMUXNet-m#5Percentage error: 2.0Params: 2.1MFLOPS: 200M
image-classification-on-cifar-10MUXNet-m#58Percentage correct: 98.0Top-1 Accuracy: 98.0
image-classification-on-cifar-100MUXNet-m#56Percentage correct: 86.1PARAMS: 2.1M
image-classification-on-imagenetMUXNet-l#852Top 1 Accuracy: 76.6%Number of params: 4.0MGFLOPs: 0.636
image-classification-on-imagenetMUXNet-m#892Top 1 Accuracy: 75.3%Number of params: 3.4MGFLOPs: 0.436
image-classification-on-imagenetMUXNet-s#943Top 1 Accuracy: 71.6%Number of params: 2.4MGFLOPs: 0.234
image-classification-on-imagenetMUXNet-xs#973Top 1 Accuracy: 66.7%Number of params: 1.8MGFLOPs: 0.132
neural-architecture-search-on-cifar-10MUXNet-m#6Top-1 Error Rate: 2.0%Parameters: 2.1MFLOPS: 200M
neural-architecture-search-on-cifar-100-1MUXNet-m#5Percentage Error: 13.9FLOPS: 200MPARAMS: 2.1M
neural-architecture-search-on-imagenetMUXNet-l#72Top-1 Error Rate: 23.4Accuracy: 76.6Params: 4.0MMACs: 318M
neural-architecture-search-on-imagenetMUXNet-m#99Top-1 Error Rate: 24.7Accuracy: 75.3Params: 3.4MMACs: 218M
neural-architecture-search-on-imagenetMUXNet-s#112Top-1 Error Rate: 28.4Accuracy: 71.6Params: 2.4MMACs: 117M
neural-architecture-search-on-imagenetMUXNet-xs#115Top-1 Error Rate: 33.3Accuracy: 66.7Params: 1.8MMACs: 66M
semantic-segmentation-on-ade20kMUXNet-m + PPM#216Validation mIoU: 35.8
semantic-segmentation-on-ade20kMUXNet-m + C1#218Validation mIoU: 32.42