suntingfeng

Hi， This project contains the unet network，having model_unet.py and train_unet.py. But they still use RNN network here. Is there a training demo using unet?

训练loss一直居高不下

1

你好，我使用MIR-1k中Wavfile作为训练集，使用UndividedWavfile作为验证集，训练的epoch为100000，但loss一直居高不下，稳定在3.0左右，请问这是为什么呢？我应该修改哪些配置呢？ 谢谢！

loss appear negative during training

1

Excuse me, why does loss appear negative during training？

Error issue using BatchPredict API

1

Hi, I use BatchPredict API for batch prediction, and when the input data shape is NDList size: 1, 0 : (2, 64) int32 or NDList size: 2 0 : (64)...