I try to train your model in my device or use your pretrained model directly on Voicebank dataset.But the both performance is different from that in your paper.I can only get the Unprocessed result in your paper.Can you give some tricks to get the same result in your paper?
I try to train your model in my device or use your pretrained model directly on Voicebank dataset.But the both performance is different from that in your paper.I can only get the Unprocessed result in your paper.Can you give some tricks to get the same result in your paper?