MLM training configurations of the text rescoring module across different data sets
| Data set | AISHELL-1 | TEDLIUM-2 | LibriSpeech |
|---|---|---|---|
| Epochs | 10 | 10 | 10 |
| Batch size | 256 | 128 | 128 |
| Gradient accumulation | 1 | 2 | 2 |
| Learning rate | 1e-5 | 1e-5 | 1e-5 |
| Warm up Ratio | 0.01 | 0.01 | 0.01 |
| Data set | AISHELL-1 | TEDLIUM-2 | LibriSpeech |
|---|---|---|---|
| Epochs | 10 | 10 | 10 |
| Batch size | 256 | 128 | 128 |
| Gradient accumulation | 1 | 2 | 2 |
| Learning rate | 1e-5 | 1e-5 | 1e-5 |
| Warm up Ratio | 0.01 | 0.01 | 0.01 |
Sharing content requires targeting cookies to be enabled. Please update your cookie preferences to use this feature.