-
Environment:
Anaconda & Python3.7
CUDA 11.2 & cudnn 8.0.5
torch-gpu 1.7.1 & torchvision 0.8.2
When I run the training command
stylegan2_pytorch --data images/128a --image-size 128 --num-trai…
-
Thank you very much for your excellent work. I only have one RTX3090. How should I modify the configuration file to achieve the model trained by you on multiple cards?
-
TensorRT-8.6.1.6+cuda12.1
GPU RTX3090 24G
-
## Problem
It is hard for a non-technical user to deploy a subnet validator, Akash offer an open source template service to do this: https://console.akash.network/templates?search=flock
## Cont…
-
## Problem
It is hard for a non-technical user to deploy a subnet validator, Akash offer an open source template service to do this: https://console.akash.network/templates?search=flock
## Cont…
-
## Problem
It is hard for a non-technical user to deploy a subnet validator, Akash offer an open source template service to do this: https://console.akash.network/templates?search=flock
## Contex…
-
Thanks for your great work, but due to my poor knowledge, what's the 'hf_token' in your code of inference? How can I get it?
-
During training mAP calculation, and even while using the `darknet detector map` command, the calculation is painfully slow. It takes several seconds to read a single image. I tried editing my `test.t…
-
1.unable to reproduce the result
MAE:0.2708086827305923,OBO:0.5
epoch:20
RTX3090+i9-12900k
2."This lightweight model trains quickly, with 15 epochs completed in 20 minutes on a GPU"
but I train…
-
您好,我在使用分布式训练的时候,并没有加快训练的速度,跟单卡训练的时间一样,请问是什么原因导致的,非常感谢,期待您的回复。