zollinger

from PIL import Image, ImageDraw
import argparse
import sys

def get_colors(image_file, numcolors=10, resize=150):
    # Resize image to speed up processing
    img = Image.open(image_file)
    img = img.copy()
    img.thumbnail((resize, resize))

blackfalcon

Git-workflow vs feature branching

When working with Git, there are two prevailing workflows are Git workflow and feature branches. IMHO, being more of a subscriber to continuous integration, I feel that the feature branch workflow is better suited, and the focus of this article.

If you are new to Git and Git-workflows, I suggest reading the atlassian.com Git Workflow article in addition to this as there is more detail there than presented here.

I admit, using Bash in the command line with the standard configuration leaves a bit to be desired when it comes to awareness of state. A tool that I suggest using follows these instructions on setting up GIT Bash autocompletion. This tool will assist you to better visualize the state of a branc

amroamroamro

Python version of the MATLAB code in this Stack Overflow post: https://stackoverflow.com/a/18648210/97160

The example shows how to determine the best-fit plane/surface (1st or higher order polynomial) over a set of three-dimensional points.

Implemented in Python + NumPy + SciPy + matplotlib.

frnsys

import numpy as np

# the 2d array of our samples,
# each component is a category label
a = np.array([[1,2,3],[4,5,6]])

# the 3d array that will be the one-hot representation
# a.max() + 1 is the number of labels we have
b = np.zeros((a.shape[0], a.shape[1], a.max() + 1))

shagunsodhani

Curriculum Learning

Introduction

• Curriculum Learning - When training machine learning models, start with easier subtasks and gradually increase the difficulty level of the tasks.
• Motivation comes from the observation that humans and animals seem to learn better when trained with a curriculum like a strategy.
• Link to the paper.

Contributions of the paper

oeway

import numpy as np
import scipy
import scipy.ndimage
from scipy.ndimage.filters import gaussian_filter
from scipy.ndimage.interpolation import map_coordinates
import collections
from PIL import Image
import numbers

__author__ = "Wei OUYANG"

melgor

import tensorflow as tf
from tensorflow.contrib.layers.python.layers import initializers
slim = tf.contrib.slim

'''
============================================================================
LinkNet: Exploiting Encoder Representations for Efficient Semantic Segmentation
============================================================================
Based on the paper: https://arxiv.org/pdf/1707.03718.pdf
'''

peteflorence

Here's a simple implementation of bilinear interpolation on tensors using PyTorch.

I wrote this up since I ended up learning a lot about options for interpolation in both the numpy and PyTorch ecosystems. More generally than just interpolation, too, it's also a nice case study in how PyTorch magically can put very numpy-like code on the GPU (and by the way, do autodiff for you too).

For interpolation in PyTorch, this open issue calls for more interpolation features. There is now a nn.functional.grid_sample() feature but at least at first this didn't look like what I needed (but we'll come back to this later).

In particular I wanted to take an image, W x H x C, and sample it many times at different random locations. Note also that this is different than upsampling which exhaustively samples and also doesn't give us fle

Tushar-N

import torch
import torch.nn as nn
import torch.nn.functional as F

import torchvision.models as tmodels
from functools import partial
import collections

# dummy data: 10 batches of images with batch size 16
dataset = [torch.rand(16,3,224,224).cuda() for _ in range(10)]

sbarratt

def get_jacobian(net, x, noutputs):
    x = x.squeeze()
    n = x.size()[0]
    x = x.repeat(noutputs, 1)
    x.requires_grad_(True)
    y = net(x)
    y.backward(torch.eye(noutputs))
    return x.grad.data