I'm surprised that they went with AMD and not an ARM vendor. ARM is pushing into the ML space big time with ARM GPU with even a Caffe fork CaffeALC on the ARM G

I'm surprised that they went with AMD and not an ARM vendor. ARM is pushing into the ML space big time with ARM GPU with even a Caffe fork CaffeALC on the ARM GPU. These GPUs are built with mobility and energy conservation, so these processors are somewhat of a natural fit for TSLA. https://github.com/OAID/caffeOnACL https://github.com/OAID/caffeOnACL