GreedyNAS: Towards Fast One-Shot NAS with Greedy Supernet

Benchmark Model Rank Results
image-classification-on-imagenetGreedyNAS-A–Top 1 Accuracy: 77.1%Number of params: 6.5MGFLOPs: 0.366
image-classification-on-imagenetGreedyNAS-B–Top 1 Accuracy: 76.8%Number of params: 5.2MGFLOPs: 0.324
image-classification-on-imagenetGreedyNAS-C–Top 1 Accuracy: 76.2%Number of params: 4.7MGFLOPs: 0.284
neural-architecture-search-on-imagenetGreedyNAS-A–Top-1 Error Rate: 22.9Accuracy: 77.1Params: 6.5M
neural-architecture-search-on-imagenetGreedyNAS-B–Top-1 Error Rate: 23.2Accuracy: 76.8Params: 5.2M
neural-architecture-search-on-imagenetGreedyNAS-C–Top-1 Error Rate: 23.8Accuracy: 76.2Params: 4.7M