On Wed, 15 Apr 2020 17:20:42 -0700, you wrote: >It seems DragonFly is practically locked out of TensorFlow. Bazel is just >too finicky.
Depending on what you want to do, and how serious any Tensorflow/ML stuff is, none of the BSD's are great. You really need GPU support to effectively do ML, which for most cases means CUDA (not available on BSD), or to a lesser extent whatever AMD uses on their GPUs. The AMD stuff is less supported in the ML world, and not sure what level of support there is in the AMD open source stuff on the BSD's regarding OpenCL, SYCL, and ROCm.
