Model
NVIDIA Deep Learning Examples
Waveglow TensorRT model (AMP, 256ch)Waveglow TensorRT model with 256-channel WN base trained with AMP
19.10.0
Copied!
1 Version
19.10.0Selected08/21/2020 2:28 AM UTC369.35 MBAccuracy: 01950 EpochsBatch Size: 0GPU: V100 Copied!
19.10.0Selected
08/21/2020 2:28 AM UTC369.35 MBAccuracy: 01950 EpochsBatch Size: 0GPU: V100
Copied!
architecture
| Key | Value |
|---|---|
| config | default |
performance
| Key | Value |
|---|---|
| quality | subjective |
training
| Key | Value |
|---|---|
| bs | 12 |
| dataset | LJSpeech-1.1 |
| epochs | 1950 |
| gpu | 8xV100 |
| training_precision | AMP |
| wn-channels | 256 |