Followups for hub_utils test PR

Created by: suchenzang

Request to swap out the binary files for something text-based / diff-able from @stephenroller
Questions around needing criterion & specify_arch flags from @lilisierrayu
Add test for decoded string (not just index tokens) from @lilisierrayu
Request to add more variations of tests (parameterized by batch size, temps, topps, etc.) + generation tasks (greedy decoding, sampling, etc.) from @lilisierrayu