-
Hello,
I'm running the benchmarking tools on BERT. For the sequence length = [1,2,4,8,64,128,384], it worked well. However, if I choose sequence length = [512, 1024, 4096], it failed even on the b…
-
I test every new version (v39.3)
The 3080TI does not manage more than 85Mhs at 280W.
Yes the Mhs and watts go down and back up.
So the graphics cards are not usable for ETH.
For comparison the n…
-
## ❓ RTX2080s better than RTX3080?
### I used RTX2080s and RTX3080 to perform semantic segmentation inference operations in the same environment(see below,also same model & same data) separately. Th…
-
你好,首先你的代码对我了解模型剪枝很有帮助,表示感谢!在剪枝剪层之后微调训练的时候出现了一个错误
![捕获0](https://user-images.githubusercontent.com/55401742/102183370-b1210c80-3ee8-11eb-8a84-b8d4915609d2.PNG)
![捕获1](https://user-images.githubuserco…
-
Dear author,
When I use RTX 3090, the PyTorch installation will come to an error with 'pip3 install -r requirements.txt'.
3090 is only adapted to the Cuda version above 11.0 while 'pip3 install…
-
6g的显存完全可以胜任推理任务,但是加了DAIN后就炸显存了,换了10G的机子还是炸
想知道作者用的配置是多大显存进行推理的?
6g of video memory is fully capable of reasoning tasks, but after the addition of DAIN will blow up the memory, replaced by a 10G ma…
-
![Screenshot from 2021-01-30 10-04-52](https://user-images.githubusercontent.com/24536554/106343895-a7116c80-62e2-11eb-87c8-8f14d39db9c9.png)
-
https://youtu.be/2-7Qn6X42nY
If you try to recreate things like that you need to do it in right circumstances. It will never happen on empty map with just few enemies. You need to put enough stres…
-
在ubuntu中编译好NCNN好尝试测试examples中YOLOX,但是速度非常慢,只有每秒一两帧,请问这是什么原因呢? 非常感谢~
-
when I follow the [automatic_mask_generator_example](https://github.com/facebookresearch/segment-anything/blob/main/notebooks/automatic_mask_generator_example.ipynb) to generating masks, It works in m…