There are a set of torchbench models that are closely traceked by project Bluberries. We would like to populate the CSV information in the HUD with this model set in order to track their performance over time as seprate column in https://hud.pytorch.org/benchmark/compilers