Summary: 35 instances, 27 unique

Text	Count
# TODO: Move this logic inside	1
# TODO: Remove after transformers package upgrade to 2.5	3
# TODO: Log a warning here	1
# TODO: Fix later when we move to proper standardized features	1
# TODO: Deprecate this method and move configuration updates directly to processors	4
# TODO: Update	2
# TODO: Later test if this can be removed	1
# TODO remove OCR order vectors; they are not needed	1
# TODO: Do clean implementation without Sequential	1
# TODO: Later support multihead	2
# TODO: To remove technical debt, a possible solution is to use	1
# TODO: we want to make the padding_idx==0, however, with custom initilization,	1
# TODO: Deprecate support for this	1
# TODO: Update after implementing decoder	1
# TODO: Update kwargs with defaults	1
# TODO possibly replace it with another sample list	1
# TODO: Make the data update generic for any type of model	1
# TODO: Clean up VizWiz IMDB from copy tokens	1
# TODO : Switch to AutoModelForPreTraining after transformers	1
# TODO: Fix question model retrieval	1
# TODO: there may still be some instability in the exponent calculation.	2
# NOTE, TODO: Code duplication w.r.t to STVQA, revisit	1
# TODO: Once omegaconf fixes int keys issue, bring this back	1
# TODO: Later upload visual dialog features as well	1
# TODO: Add Help flag here describing MMF Configuration	1
# TODO: Move this logic inside	1
TODO: Update on docs sprint	1