UniNet: Unified Architecture Search with Convolution, Transformer, and MLP

Open paper Other version
Benchmark Model Rank Results
neural-architecture-search-on-imagenetUniNet-B0#11Top-1 Error Rate: 19.2Params: 11.5MFLOPs: 555M
image-classification-on-imagenetUniNet-B6#96Top 1 Accuracy: 87.4%Number of params: 117MGFLOPs: 51
image-classification-on-imagenetUniNet-B5#120Top 1 Accuracy: 87%Number of params: 72.9MGFLOPs: 20.4
image-classification-on-imagenetUniNet-B0#723Top 1 Accuracy: 80.8%Number of params: 11.5MGFLOPs: 0.555
image-classification-on-imagenetUniNet-B0–Top 1 Accuracy: 79.1%Number of params: 11.9MGFLOPs: 0.56
image-classification-on-imagenetUniNet-B1–Top 1 Accuracy: 80.4%Number of params: 14MGFLOPs: 0.99
image-classification-on-imagenetUniNet-B2–Top 1 Accuracy: 82.7%Number of params: 22.5MGFLOPs: 2.4
image-classification-on-imagenetUniNet-B4–Top 1 Accuracy: 84.2%Number of params: 73.5MGFLOPs: 9.9
image-classification-on-imagenetUniNet-B5–Top 1 Accuracy: 85.2%Number of params: 73.5MGFLOPs: 23.2

Rank counts only results with a code link.