[WIP][Do Not Review] enabling tests that depends on downloading from google drive

[parmeet]

Enabling back all the tests that depend on downloading data from google drive (essentially reversing what was done in #1171)

Our current understanding is that based on:

1. we are caching .data folder (where all the data downloaded goes during testing) which is refreshed every week.
2. Based on merge of [BC-breaking] Split raw sequence tagging datasets into individual files #1176, all the google drive data now have defined path in .data folder (defined using _PATH variable)
3. If data is available in .data folder, it will used during testing (and will not trigger the download: refer

   text/torchtext/utils.py

   Line 79 in 1197514

   return path

   )

Ideally this should ensure that we do not download if the data is already cached by CCI. Any blind spot here?

[parmeet]

Closing it up, No longer relevant at the moment.