Update README.md
Browse files
README.md
CHANGED
|
@@ -94,7 +94,6 @@ The following table describes the available features. All features consists of t
|
|
| 94 |
| pool | 3072 | Pooled statistics before the bottleneck speaker embedding layer, extracted before ReLU layer.
|
| 95 |
| attention | 3072 | Same as the pooled statistics but with the attention weights applied.
|
| 96 |
| embedding | 192 | The standard ECAPA2 speaker embedding.
|
| 97 |
-
|
| 98 |
<!--
|
| 99 |
The following table describes the available features:
|
| 100 |
|
|
@@ -104,7 +103,6 @@ The following table describes the available features:
|
|
| 104 |
| Global Feature | Uniform effective receptive field of each frame-level feature in the frequency dimension.| Generally capture intra-speaker variance better then speaker embeddings. E.g. speaker profiling, emotion recognition. | gfe1, gfe2, gfe3, pool
|
| 105 |
| Speaker Embedding | Uniform effective receptive field of each frame-level feature in the frequency dimension.| Best for tasks directly depending on the speaker identity (as opposed to speaker characteristics). E.g. speaker verification, speaker diarization. | embedding
|
| 106 |
-->
|
| 107 |
-
|
| 108 |
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 109 |
|
| 110 |
## Citation
|
|
|
|
| 94 |
| pool | 3072 | Pooled statistics before the bottleneck speaker embedding layer, extracted before ReLU layer.
|
| 95 |
| attention | 3072 | Same as the pooled statistics but with the attention weights applied.
|
| 96 |
| embedding | 192 | The standard ECAPA2 speaker embedding.
|
|
|
|
| 97 |
<!--
|
| 98 |
The following table describes the available features:
|
| 99 |
|
|
|
|
| 103 |
| Global Feature | Uniform effective receptive field of each frame-level feature in the frequency dimension.| Generally capture intra-speaker variance better then speaker embeddings. E.g. speaker profiling, emotion recognition. | gfe1, gfe2, gfe3, pool
|
| 104 |
| Speaker Embedding | Uniform effective receptive field of each frame-level feature in the frequency dimension.| Best for tasks directly depending on the speaker identity (as opposed to speaker characteristics). E.g. speaker verification, speaker diarization. | embedding
|
| 105 |
-->
|
|
|
|
| 106 |
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
|
| 107 |
|
| 108 |
## Citation
|