Dzmitry Bahdanau
Dzmitry Bahdanau
Tests must be broken now, this should fix them.
We should have a ready demo to reproduce results from http://karpathy.github.io/2015/05/21/rnn-effectiveness/ This is mostly matter of changing the input in Markov chain examples. And also we don't have deep RNNs...
When I unpickle an iterator `iter` over an out-of-memory H5PyDataset which was pickled in the middle of epoch, I can't do `next(iter)`. In fact I have already traced the problem:...
I would be useful it have a sourcewise transformer that would apply a mapping to `which_sources`. I am going to code such thing soon, this issue is a reminder for...
We should consider including "validation" split to our MNIST dataset. It has certain official status, as far as I know.
For example, for `SourcewiseTransformer`.
Thanks for releasing code! I'm trying to understand how this implementation handles multi-value slot values introduced in MultiWOZ 2.1 (such as e.g. cheap|moderate). It appears that this implementation considers a...
And Python 2 has reached the end of line this year.
It appears that recent changes that were applied to the `tables.json` of SPIDER dataset were not propagated to this repository. My understanding is that these changes would be desirable for...