zheng-xing comments

Repositories
Issues
Comments

Results 3 comments of


                                            zheng-xing

LOSS value can not be reduced during training

I have the same problem in my case, where I need to segment two foreground class from the background. The dice loss decreases to around 0.67. For me, the problem...

inference problom about energy and pitch

> hi ming024 i have some questions about inference。 1）energy and pitch change ,but the generat wav is similiar. i set pitch 0.7、1.0、5， the wav sounds similiar. "--pitch_control", type=float, default=5.0,...

No significant performance improvement on UNET semantic segmentation model.

> Hi, > I used TFTRT to optimize the 3D UNET segmentation. I couldn't find any significant improvement in the speed. > On a P100 GPU, for a batch size...