scan implements strict semantics so it will always execute the same number of timesteps no matter what the accumulator is (nan).
while_loop implements dynamic execution (quit once cond is not met) and at the same time allows parallel execution when some ops are not dependent on accumulator.
If you read the code for `dynamic_rnn` and contrib.legacy Seq2seq model you'll find while_loop. I have yet to see tensorflow library code using tf.scan anywhere!
Also, internally, scan is defined using while_loop. In my code, I find scan lacking in RNN and always have to fall back to while_loop.
Here is video of a talk by the RNN/Seq2Seq author himself:
I read the article and seems to be well-written though lacking.
For even more customized RNNs such as attention mechanism, beam search as in Seq2Seq, you'll need to skip the tf.dynamic_rnn abstraction and use a symbolic loop directly: tf.while_loop
YC partner Justin Kan has a sketch of magic mushrooms framed in his living room. I saw it on his Snapchat when he showed his followers around his house.