Multi-object Classification on MS-COCO
66.92Class-wise F1DiCENet
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DiCENet# Params=5.1 M, # FLOPs=298 M, fine-tuning=true2019.06 | 66.92 | 73.41 | |
| ESPNetv2# Params=3.5 M, # FLOPs=284 M, fine-tuning=true2019.06 | 63.41 | 69.23 | |
| ShuffleNetv2# Params=3.5 M, # FLOPs=300 M, fine-tuning=true2019.06 | 60.42 | 67.58 |