[FastPitch] unused parameters in attention module

NVIDIA / DeepLearningExamples

State-of-the-Art Deep Learning scripts organized by models - easy to train and deploy with reproducible accuracy and performance on enterprise-grade infrastructure.

12.94k stars 3.12k forks source link

[FastPitch] unused parameters in attention module #1310

Open roedoejet opened 1 year ago

roedoejet commented 1 year ago

Related to Model/Framework(s) SpeechSynthesis/FastPitch

Describe the bug There is a convolutional layer attn_proj that is declared, but unused anywhere which results in unused parameters in the model. I don't see evidence of this layer in the related paper, so maybe it should be removed?