Research Article
A Fast Learning Method for Multilayer Perceptrons in Automatic Speech Recognition Systems
Table 1
WERs and model scales.
| | Task | NUM-PARA () | WER |
| | Conventional | ≈6.64 | 16.32% | | PA-SVD-384 | ≈4.74 | 16.48% | | PA-SVD-256 | ≈3.31 | 16.55% | | PA-SVD-128 | ≈1.88 | 16.71% |
|
|
