Summary: 22 instances, 20 unique

Text	Count
# FIXME: GradientAttribution is provided as a string due to a circular import.	1
# TODO support multiple baselines	1
# TODO: Output widget only captures beginning of server logs. It seems	1
TODO: Either restore model state after done (would have to place functionality	1
help_info: Optional[str] = None  # TODO fill out help for each method	1
# TODO: restore prior state	1
# TODO Reconsider this check if _batched_generator is used for non gradient-based	1
TODO: For models that can have tuples as activations, we should	1
//TODO: Refactor the visualization table as a <table> instead of columns, in order to have cleaner styling	1
# TODO:	1
TODO: Support for checkpoint type. Currently only supports model parameters as saved	2
# TODO: allow loss_fn to be Callable	1
# TODO support batch size > 1	1
#TODO: Remove when bugs are fixed	1
r"""TODO:	1
# TODO: allow loss_fn to be Callable	1
TODO: we can potentally also output tuples of attributions. This might be	1
# TODO find a better way of checking if a module is a container or not	1
# FIXME it look like it is very difficult to make torch.normal	1
):  # TODO: allow loss_fn to be Callable	2