add LFMMI loss #1725
Merged
add LFMMI loss #1725
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
robin1001
reviewed
Mar 9, 2023
wenet/transformer/asr_model.py
Outdated
| @@ -89,6 +94,9 @@ def forward( | |||
| text: (Batch, Length) | |||
| text_lengths: (Batch,) | |||
| """ | |||
| if self.lfmmi_dir != '': | |||
There was a problem hiding this comment.
Why load it in forward? I think we should load it in construction.
There was a problem hiding this comment.
Then there is no need to use hasattr when loading resource
There was a problem hiding this comment.
Why load it in forward? I think we should load it in construction.
because i need to decorate it with torch.jit.ignore, i dont want to decorate whole forward function with torch.jit.ignore
There was a problem hiding this comment.
now we move load_mmi_resource into init and remove hasattr judgement
robin1001
reviewed
Mar 9, 2023
| @torch.jit.ignore(drop=True) | ||
| def _calc_lfmmi_loss(self, encoder_out, encoder_mask, text): | ||
| ctc_probs = self.ctc.log_softmax(encoder_out) | ||
| supervision_segments = torch.stack( |
There was a problem hiding this comment.
why we should move it to cpu?
There was a problem hiding this comment.
k2 requires supervision_segments on cpu
robin1001
approved these changes
Mar 10, 2023
|
@aluminumbox 请问现在lfmmi在runtime推理中能够使用吗 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
add LFMMI loss, add torch.jit.ignore for lfmmi function