Human Activity Recognition on HAR-2 (test)
97.2Test AccuracycoRNN
Evaluation results
| Method | Links | |
|---|---|---|
| coRNN# units=64, # params=9k2020.10 | 97.2 | |
| tiny coRNN# units=20, # params=1k2020.10 | 96.5 | |
| incremental RNN# units=64, # params=4k2020.10 | 96.3 | |
| FastGRNN# units=80, # params=7k2020.10 | 95.6 | |
| FastRNN# units=80, # params=7k2020.10 | 94.5 | |
| LSTM# units=64, # params=16k2020.10 | 93.7 | |
| GRU# units=75, # params=19k2020.10 | 93.6 | |
| anti.sym. RNN# units=120, # params=8k2020.10 | 93.2 |