zmtttt

have you sloved ? I met the same problem

"Hello, this is my loss graph. It seems that there isn't a competitive relationship forming. Could you suggest some measures for improvement?" thanks！！！！  @Yu-Chen-Yi @xulin23 @Orrinn @hzyliusha @aliinassiri

"Hello, I'd like to ask, what are the final losses for the generator and the discriminator?" @xulin23 @erf1122

是onnx2trt_ptq这个

result = self.forward(*input, **kwargs) TypeError: Segnet.forward() missing 2 required positional arguments: 'pix_T_cams' and 'cam0_T_camXs' why？？ my_onnx_inputs include those, torch.Size([4, 6, 3, 900, 1600]) torch.Size([4, 6, 4, 4]) torch.Size([4, 6, 4,...

This revised statement clearly communicates that you have made improvements to the script, but there are still problems with the captured timing data for rank0 and rank2.... like this: the...

> Thank you @zmtttt for your interest in veScale! If I understand your question correctly, you expressed concern about the visible gaps in profiled overheads between the first few iterations...

Hi! (1)I've set up my experiment again with step=30 and batch=128. I visualized the JSON files generated at steps 2, 20, and 30. (2)I noticed that as the step increases,...

> > Hi! (1)I've set up my experiment again with step=30 and batch=128. I visualized the JSON files generated at steps 2, 20, and 30. (2)I noticed that as the...

> > > > Hi! (1)I've set up my experiment again with step=30 and batch=128. I visualized the JSON files generated at steps 2, 20, and 30. (2)I noticed that...