UserWarning: Converting sparse IndexedSlices to a dense Tensor of unknown shape. This may consume a large amount of memory. #27
Description
When I running the train_and_eval.py on CNNDailyMail dataset, it often has the problem'UserWarning: Converting sparse IndexedSlices to a dense Tensor of unknown shape. This may consume a large amount of memory.'
I have found that the dataset.tensorshape have been defineds in advance as [None], and it's feeded into the trainning. Maybe it's the reason.
Can anybody give a solution？ Thanks!