https://instinct.radeon.com/en/6-deep-learning-projects-amd-radeon-instinct/<p>https://rocm.github.io/<p>Despite the ridiculous price point, I enjoy the portability and battery life of MacBooks. Almost all of my ML projects are on AWS or Linode or something, but if I were to even consider forking over for a MacBook Pro w/ a GPU, some sort of GPU ML support would be a must.<p>Anyone have experience with Radeon deep learning libraries? Thoughts?<p>(not trying to have a debate about the obvious impracticalities of mac laptops – just curious if anyone has done ML w/ Radeon libraries)
Maybe you found these StackOverflow Threads and Websites already
<a href="https://stackoverflow.com/questions/37892784/using-keras-tensorflow-with-amd-gpu/40814520" rel="nofollow">https://stackoverflow.com/questions/37892784/using-keras-ten...</a><p><a href="https://en.wikipedia.org/wiki/Comparison_of_deep_learning_software" rel="nofollow">https://en.wikipedia.org/wiki/Comparison_of_deep_learning_so...</a><p><a href="https://gpuopen.com/compute-product/miopen/" rel="nofollow">https://gpuopen.com/compute-product/miopen/</a><p>From reading these articles the consensus is that if need speed now and can't wait use an NVIDIA while AMD plays catchup. Who knows how long this situation will last.