GPU is not used while training

[Isdriai]

Hi,

I followed the doc to use MXnet on my GPU (RTX 3080 laptop).

But when I debug I found the CPU as available device (as you can see here

In my file build.gradle.kts I put these entries in my dependencies:

implementation("ai.djl.mxnet:mxnet-engine:0.18.0")

// https://mvnrepository.com/artifact/ai.djl.mxnet/mxnet-native-cu102mkl
implementation("ai.djl.mxnet:mxnet-native-cu102mkl:1.9.1")

So I'm wondering if the problem is my CUDA version

nvcc --version                                                                                                                                                                                                              
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2022 NVIDIA Corporation
Built on Wed_Jun__8_16:49:14_PDT_2022
Cuda compilation tools, release 11.7, V11.7.99
Build cuda_11.7.r11.7/compiler.31442593_0

Maybe the dependency doesn't exist for my CUDA version, but I'm not sure.

I have to downgrade my CUDA version ?

[frankfliu]

1. MXNet must use matching cuda version, your cuda is 11.7, you can try downgrade to CUDA 11.2 and use MXNet 1.9.1 (ai.djl.mxnet:mxnet-native-cu112mkl:1.9.1:linux-x86_64)

According to this, GeForce RTX 3080's compute capability is 8.6, I'm not sure if MXNet 1.9.1 support 8.6