mkolod

import mesosphere.mesos.util.FrameworkInfo
import org.apache.mesos.MesosSchedulerDriver


/**
 * @author Tobi Knaup
 */

object Main extends App {

mkolod

mkolod

mkolod

# Working example for my blog post at:
# http://danijar.com/introduction-to-recurrent-networks-in-tensorflow/
import functools
import sets
import tensorflow as tf
from tensorflow.models.rnn import rnn_cell
from tensorflow.models.rnn import rnn


def lazy_property(function):

mkolod

from keras.layers import GRU, initializations, K

class GRULN(GRU):
    '''Gated Recurrent Unit with Layer Normalization
    Current impelemtation only works with consume_less = 'gpu' which is already
    set.

    # Arguments
        output_dim: dimension of the internal projections and the final output.
        ...: see GRU documentation for all other arguments.

mkolod

Scala's "type T in class C as seen from a prefix type S" with examples

Introduction

Recently, I found myself in need to precisely understand Scala's core typechecking rules. I was particulary interested in understanding rules responsible for typechecking signatures of members defined in classes (and all types derived from them). Scala Language Specification (SLS) contains definition of the rules but lacks any examples. The definition of the rules uses mutual recursion and nested switch-like constructs that make it hard to follow. I've written down examples together with explanation how specific set of rules (grouped thematically) is applied. These notes helped me gain confidence that I fully understand Scala's core typechecking algorithm.

As Seen From

Let's quote the Scala spec for As Seen From (ASF) rules numbered for an easier reference:

mkolod

import numpy as np
cimport numpy as np

cdef extern from "numpy/arrayobject.h":
    void PyArray_ENABLEFLAGS(np.ndarray arr, int flags)
    object PyArray_SimpleNewFromData(int nd, np.npy_intp* dims, int typenum, void* data)
    void _import_array()

cdef np.ndarray[np.int32_t, ndim=1, mode='c'] from_pointer_to_ndarray(np.npy_intp n_elements,
                                                                      np.int32_t* values):

mkolod

/* https://spectreattack.com/spectre.pdf */
#include <stdio.h>
#include <stdlib.h>
#include <stdint.h>
#ifdef _MSC_VER
#include <intrin.h> /* for rdtscp and clflush */
#pragma optimize("gt",on)
#else
#include <x86intrin.h> /* for rdtscp and clflush */
#endif

mkolod

# -*- mode: dockerfile -*-
# Work in progress, some of the manual steps below will be fixed in a subsequent release.
# Dockerfile to build libmxnet.so, and a python wheel for the Jetson TX1 and TX2
# Builds from Github MXNet master branch
# Once complete copy artifacts from /work/build to target device.
# Install by running 'pip wheel name_of_wheel.whl' and copying the .so to a folder on your LD_LIBRARY_PATH

FROM nvidia/cuda:8.0-cudnn5-devel as cudabuilder

FROM dockcross/linux-arm64

mkolod

// Just before switching jobs:
// Add one of these.
// Preferably into the same commit where you do a large merge.
//
// This started as a tweet with a joke of "C++ pro-tip: #define private public",
// and then it quickly escalated into more and more evil suggestions.
// I've tried to capture interesting suggestions here.
//
// Contributors: @r2d2rigo, @joeldevahl, @msinilo, @_Humus_,
// @YuriyODonnell, @rygorous, @cmuratori, @mike_acton, @grumpygiant,