In this video I’m showing off DirectML, a tool made by Microsoft that let’s you use almost any GPU for machine learning acceleration. I show how to get it running, using an AMD GPU as the example. I also show how it makes my 3070ti slower than my 3060, so there’s times not to use it.
Commands I use: curl https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh \ -o Miniconda3-latest-Linux-x86_64.sh
bash Miniconda3-latest-Linux-x86_64.sh
conda create -n tensorflow-directml python=3.6
conda activate tensorflow-directml
pip install tensorflow-directml
python
import tensorflow as tf
print(tf.test.is_gpu_available())
