Alexander Botev

I just want to point out for anyone using the code that you are approximating the empirical Fisher, not the true Fisher matrix, thus the method is not exactly the...

What I mean is that when we have a very long one liner: ``` rust try!(self.nodes.get(index).ok_or_else(|| "The index is out of bounds")).as_ref().ok_or_else(|| format!("The requested node {} is None", index)) ```...

Error from pip install

10

Hi, so recently trying to install from `pip install` when I try running tensorboard from the command line I get: ``` RuntimeError: module compiled against API version 0xb but this...

Added fixes to problems that occur on python3

2

The following errors happened when initially running tensorboard in python3: 1. The `StringIO` was giving the following error: TypeError: string argument expected, got 'bytes' Applied the fix from here: https://stackoverflow.com/questions/6479100/python-stringio-replacement-that-works-with-bytes-instead-of-strings...

Do you guys have plans on implementing the merge summaries bit?

FlowDistribution

4

Following the conversation in #70

Exponential families and natural parameters

9

It would really great if there is a way of constructing an easy way of switching between standard and natural parameters. This particularly is to address more advanced techniques such...

Add Serialize/Deserialize for Duration

35

For features `serde` and `rustc-serialize`

Batch normalization - why not use theano?

11

So, is there any reason why Lasagne has still not started using the `T.nnet.batch_normalization_train/test` for the `BatchNormalizationLayer`? I did a small test and with `fast_compile` the Lasagne implementation is very...

Dynamic shapes

44

Reading the current spec it seems that NNEF supports only fixed (constant) shape sizes. Is that correctly understood or not and if so do you intend to support dynamic/variable shapes?