Ilya Kavalerov

Adds WideResnet at sizes 48, 96. Adds dataloading for entire STL10 dataset. Moves transforms to `dataset/` code. Adds results and instructions to `README.md`.

Benchmark with picovoice

6

Have you considered adding your engine as a comparison in: https://github.com/Picovoice/wake-word-benchmark ?

Adds training script

2

Hi and thank you for sharing the DPTNet code. I've added a training script that you will recognize from: https://github.com/ShiZiqiang/dual-path-RNNs-DPRNNs-based-speech-separation The `solver.py` is basically copy-pasted from that repo, as well...

Since this is an NP hard problem formulatable as [3-dimensional / tri-partite matching](http://en.wikipedia.org/wiki/3-dimensional_matching#CITEREFCrescenziKannHalld.C3.B3rssonKarpinski2000) (MAX 3DM) or partition into triangles (PIT), to be correct is to use brute force search.

Here is an algorithm that generates triplets in a better way: for simplicity, start with K=1, N=10, f(n)=n, g(q,w)=w/q. 1. construct an undirected graph - connect people who do not...

When phrased as the optimization problem [here](https://en.wikipedia.org/wiki/3-dimensional_matching#Optimization_problem) and [here](https://www.nada.kth.se/~viggo/wwwcompendium/node143.html), there are 2 well know approaches. The first gets a round of sups that is within 1/3 of the best possible...

What is up with the transaction ids like `AtKc7wqJuLy/VXGIR1MxVo46zOGrML+HwFDmiTv137Op` instead of `3JuETz96LrotPgv9db7awdfENVW4p149HgEjQ33wtNxM7o9gLF8Q53`? Looking at this [block](https://www.bitcloutpulse.com/explorer/blocks/000000000000bb022da16a8b0990995b05431cc4ddbd9977278cec5dea3d1e78?search=000000000000bb022da16a8b0990995b05431cc4ddbd9977278cec5dea3d1e78), its first transaction should be `3JuETz96LrotPgv9db7awdfENVW4p149HgEjQ33wtNxM7o9gLF8Q53` but it appears as `AtKc7wqJuLy/VXGIR1MxVo46zOGrML+HwFDmiTv137Op` when I do...