catching dataset-specific metrics

making small text changes

adding the catch for glue and paws (for loading them) -- thanks @quentin !

adding more emoji

adding little fixes

adding some fixes (paw still isn't working though)

adding some hacky task type detection

hardcoding metrics (for now)

Oops, adding requirements.txt

Initial version of the Evaluation Buddy -- currently most things are hardcoded (e.g. the dataset list), but the goal it to make it all compatible with the Hub!

Update README.md

initial commit