DeLesley Hutchins

Layers can be reused, but blocks cannot -- a block is hooked into its surrounding context in various ways. So just do the following: word_embedding = Embedding(100,100,name="word2vec") def word2vec(): return...

No existing code. All of our models have been over trees or DAGs. On Tue, May 30, 2017 at 2:42 AM, bdqnghi wrote: > Hi, I would like to ask...

Honestly, I myself am not going to get to it any time soon. Someone else want to try? On Tue, Jun 6, 2017 at 12:01 PM, bdqnghi wrote: > @delesley...

The warning about gradients_impl.py is always there; ignore it. That's tensorflow complaining that we backpropagated through tf.gather. Your code seems valid -- there's nothing that jumps out at me. If...

In theory I'm sure it's possible, but this is not something I've tried to do, so I don't have any instructions. You'll have to build from source, and it will...

As you have noticed, the Tensorflow Fold repository is not being actively updated right now, in part because there are no users. :-) In the past, when I pushed an...

TensorFlow Fold can certainly help with recursive neural networks. You simply write a recursive python function, and it will handle the autobatching. The library that you want to use is...

You can use a metric to collect information from every node. You can't use the metric as an input to a block, but you can access it as an output...

Metrics are really intended for things like loss or perplexity, where you can sum all the outputs over a batch, without worrying about which node in the tree the value...

You can't mix TensorFlow expressions like "tf.constant", and block combinators like that. Block combinators operate on blocks, not tensors. TensorFlow code has to be wrapped in td.FromTensor or td.Function. On...