Conversation
Introduced in #42806
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
zucchini-nlp
left a comment
zucchini-nlp
left a comment
There was a problem hiding this comment.
Yeah, I forgot to check if _processor_class is used in other places. Not sure if the current fix works because feature extractor type is not the same as processor class. I also had a PR in #42920, just waiting for a review 😄
| feature_extractor_type = feature_extractor.to_dict().get("feature_extractor_type", None) | ||
| if ( | ||
| feature_extractor._processor_class | ||
| and feature_extractor._processor_class.endswith("WithLM") | ||
| feature_extractor_type | ||
| and feature_extractor_type.endswith("WithLM") |
There was a problem hiding this comment.
I think the feature extractor is a simple Wav2Vec2FeatureExtractor so this might not work. I found only one processor with LM in hte codebase tbh
There was a problem hiding this comment.
Yeah, that was my question too, I wasn't sure how many cases would be impacted. Closing this in favor of yours then.
|
Closing in favor of #42920. |
3 participants
Introduced in #42806
Before this PR, the following failed: