modify doc

This commit is contained in:
yaozengwei 2022-07-07 11:25:52 +08:00
parent f9c6014579
commit 12c176c443

View File

@ -1138,6 +1138,9 @@ class EmformerEncoder(nn.Module):
Streaming Speech Recognition*
[:footcite:`shi2021emformer`].
In this model, the memory bank computation is simplifed, using the averaged
value of each chunk as its memory vector.
Args:
d_model (int):
Input dimension.