mrgloom

## Convolutional Arithmetic.ipynb

      
              1 file
            
          
              0 forks
            
          
              0 comments
            
          
              0 stars
            
          
                mrgloom
                / Convolutional Arithmetic.ipynb
            
            
              Created
              May 12, 2017 16:44
                — forked from akiross/Convolutional Arithmetic.ipynb
            
              
                Few experiments on how convolution and transposed convolution (deconvolution) should work in tensorflow.
              
          
      View Convolutional Arithmetic.ipynb
  
    
      Sorry, something went wrong. Reload?
      Sorry, we cannot display this file.
      Sorry, this file is invalid so it cannot be displayed.
      
          Viewer requires iframe.
      
    
## online_stats.py
# Author: Kyle Kaster
# License: BSD 3-clause
import numpy as np


def online_stats(X):
    """
    Converted from John D. Cook
    http://www.johndcook.com/blog/standard_deviation/
    """

## VGG-16
name: "VGG_ILSVRC_16_layers"

layer {
  name: "train-data"
  type: "Data"
  top: "data"
  top: "label"
  include {
    stage: "train"
  }

## Find header and libs cheatsheet
#To find function in headers
grep -n -r <function_name> <path_to_opencv_include_folder>

#To find function in libs
nm -C -A <path_to_opencv_lib_folder>/*.so | grep <function_name> | grep -v U

#Find package installed via apt-get
apt list --installed | grep <package_name>
#Find files related to package
dpkg -L <package_name>

## detect_multiscale.cpp
#include <opencv2/opencv.hpp>
#include <vector>

using namespace cv;
using namespace std;

int main () {
  Mat img = imread("lena.jpg");
  CascadeClassifier cascade;
  if (cascade.load("haarcascade_frontalface_alt.xml")) {

## segnet_simple_train.prototxt
name: "segnet"

layer {
  name: "data"
  type: "DenseImageData"
  top: "data"
  top: "label"
  dense_image_data_param {
    source: "/SegNet/CamVid/train.txt"	# Change this to the absolute path to your data file
    batch_size: 4   			# Change this number to a batch size that will fit on your GPU

## segnet_train_downsampled.prototxt
name: "VGG_ILSVRC_16_layer"
layer {
  name: "data"
  type: "DenseImageData"
  top: "data"
  top: "label"
  dense_image_data_param {
    source: "/home/myuser/Downloads/SegNet/SegNet-Tutorial/CamVid/train.txt"	# Change this to the absolute path to your data file
    batch_size: 1   			# Change this number to a batch size that will fit on your GPU
    shuffle: true

## pegasos.py
class OnlineLearner(object):
    def __init__(self, **kwargs):
        self.last_misses = 0.
        self.iratio = 0.
        self.it = 1.
        self.l = kwargs["l"]
        self.max_ratio = -np.inf
        self.threshold = 500.

    def hinge_loss(self, vector, cls, weight):

## rbm.py
"""
Code for training RBMs with contrastive divergence. Tries to be as
quick and memory-efficient as possible while utilizing only pure Python
and NumPy.
"""

# Copyright (c) 2009, David Warde-Farley
# All rights reserved.
#
# Redistribution and use in source and binary forms, with or without

## kmtransformer.py
from sklearn.base import BaseEstimator, TransformerMixin
from sklearn.metrics.pairwise import rbf_kernel


class KMeansTransformer(BaseEstimator, TransformerMixin):
    def __init__(self, centroids):
        self.centroids = centroids

    def fit(self, X, y=None):
        return self
	# Author: Kyle Kaster
	# License: BSD 3-clause
	import numpy as np


	def online_stats(X):
	"""
	Converted from John D. Cook
	http://www.johndcook.com/blog/standard_deviation/
	"""
	name: "VGG_ILSVRC_16_layers"

	layer {
	name: "train-data"
	type: "Data"
	top: "data"
	top: "label"
	include {
	stage: "train"
	}
	#include <opencv2/opencv.hpp>
	#include <vector>

	using namespace cv;
	using namespace std;

	int main () {
	Mat img = imread("lena.jpg");
	CascadeClassifier cascade;
	if (cascade.load("haarcascade_frontalface_alt.xml")) {
	name: "segnet"

	layer {
	name: "data"
	type: "DenseImageData"
	top: "data"
	top: "label"
	dense_image_data_param {
	source: "/SegNet/CamVid/train.txt" # Change this to the absolute path to your data file
	batch_size: 4 # Change this number to a batch size that will fit on your GPU
	name: "VGG_ILSVRC_16_layer"
	layer {
	name: "data"
	type: "DenseImageData"
	top: "data"
	top: "label"
	dense_image_data_param {
	source: "/home/myuser/Downloads/SegNet/SegNet-Tutorial/CamVid/train.txt" # Change this to the absolute path to your data file
	batch_size: 1 # Change this number to a batch size that will fit on your GPU
	shuffle: true
	class OnlineLearner(object):
	def __init__(self, **kwargs):
	self.last_misses = 0.
	self.iratio = 0.
	self.it = 1.
	self.l = kwargs["l"]
	self.max_ratio = -np.inf
	self.threshold = 500.

	def hinge_loss(self, vector, cls, weight):
	"""
	Code for training RBMs with contrastive divergence. Tries to be as
	quick and memory-efficient as possible while utilizing only pure Python
	and NumPy.
	"""

	# Copyright (c) 2009, David Warde-Farley
	# All rights reserved.
	#
	# Redistribution and use in source and binary forms, with or without
	from sklearn.base import BaseEstimator, TransformerMixin
	from sklearn.metrics.pairwise import rbf_kernel


	class KMeansTransformer(BaseEstimator, TransformerMixin):
	def __init__(self, centroids):
	self.centroids = centroids

	def fit(self, X, y=None):
	return self