Table 7 Word error rates of the DNN and CNN models on the Hungarian LVCSR task
From: Phone recognition with hierarchical convolutional deep maxout networks
Network type | Development set (%) | Test set (%) |
|---|---|---|
DNN, ReLU | 17.7 | 17.0 |
DNN, maxout | 17.4 | 16.6 |
DNN, maxout, hybrid DPT | 17.2 | 16.5 |
CNN | 16.5 | 15.9 |
Hierarchical CNN | 16.1 | 15.5 |