Would it be helpful in Speech Recognition especially in acoustic model to replace CTC with ACE?

summerlvsong / Aggregation-Cross-Entropy

Aggregation Cross-Entropy for Sequence Recognition. CVPR 2019.

303 stars 60 forks source link

Would it be helpful in Speech Recognition especially in acoustic model to replace CTC with ACE? #10

Open nemoramo opened 5 years ago

nemoramo commented 5 years ago

Hello, I am writing this topic to ask if it will be useful in speech recognition tasks. I am going to test your ACE loss on my acoustic model. Hope it can produce comparable performance. I will show the result later.

summerlvsong commented 5 years ago

Yes, after feature extraction, there is no much difference between acoustic model and OCR model.