3 ms·
Macs have GPUs and their architecture means that GPU has access to the full system RAM. Cuda isn’t a requirement for running ML workloads on a GPU.
by breckenedge 3y ago
Macs have GPUs and their architecture means that GPU has access to the full system RAM. Cuda isn’t a requirement for running ML workloads on a GPU.