After https://github.com/pytorch/ao/pull/3911 is landed we should update the `use_triton_kernel` to use `nvfp4_quantize_kernel_choice` both in OSS and in internal callsites
After #3911 is landed we should update the
use_triton_kernelto usenvfp4_quantize_kernel_choiceboth in OSS and in internal callsites