Table 3 Model ablation study.
From: Adaptive context biasing in transformer-based ASR systems
Model | Dev | Test | ||
|---|---|---|---|---|
CER% | CERR% | CER% | CERR% | |
S0: Baseline | 5.92 | \(-6.47\) | 6.55 | \(-8.89\) |
S1: Baseline + Attention | 5.62 | 1.08 | 6.10 | 6.87 |
S2: S1 + Detector | 5.56 | – | 6.01 | – |
