Han Zhang

I cannot make hardware acceleration work on Jetson Nano. After turning the hardware acc to enabled, the client error message indicates there is no `libvdpau_nvidia.so`. Any ideas?

The links to training logs on ADE20K is missing. Can you provide your training log?

How do you create fig. 3?

Cannot reproduce the reported mIoU of Segformer on Cityscapes.

8

I cannot reproduce the [82.25 mIoU](https://github.com/open-mmlab/mmsegmentation/tree/master/configs/segformer#:~:text=1.39-,82.25,-83.48) of Segformer on Cityscapes at my machine. I use 8 3090, with --seed 0 and --deterministic. The log is [attached](https://github.com/open-mmlab/mmsegmentation/files/9332488/20220812_015905.log.json.txt). Any ideas? Thanks in...

Will pretraining on Object365 helps results on ADE20K to get to 60+ mIoU?

5

[Mask DINO](https://github.com/IDEACVR/MaskDINO) says they pretrained on Object365 detection dataset, and the results on semantic segmentation on ADE20k gets to 60.8, which is way higher than backbones such as convnext and...

Reproducing the "discharge" and "full" model

3

Hi, thanks for this very good and organized codebase. Following your code, the "variance" model is easy to reproduce. However, I have trouble reproducing the performance of the "discharge" and...

About the theoretical value of the GPU

2

请问沐神： 1. 在notebook中指向的wiki里，3090ti的理论值40是从表中的core boosted value（[39.997](https://en.wikipedia.org/wiki/GeForce_30_series#:~:text=33.546-,(39.997),-0.524%0A(0.625))）得到的吗？ 2. 我在自己的很多块3090上，用CUDA11.7和nvidia-driver 525跑出来的TFLOPS都只有24，距离base（29.3）和boost（35.6）的理论值都有一定的差距。请问notebook中用3090ti跑的TFLOPS是经过超频的吗？要想达到接近理论值的FLOPS需要做怎样的设置呢？