shangeth

AlexNet, A Deep Learning Breakthrough

ImageNet Classification with Deep Convolutional Neural Networks

Authors : Alex Krizhevsky, Ilya Sutskever, and Geoffrey Hinton

This was one of the most influential publication in the field of Computer Vision.This model won the ILSVRC 2012 Competition.

shangeth

import torch.nn as nn

class LRN(nn.Module):
  def __init__(self, size, alpha=1e-4, beta=0.75, k=1):
    super(LRN, self).__init__()
    self.avg = nn.AvgPool3d(kernel_size =(size,1,1), stride=1, padding=int((size-1)/2))
    self.alpha = alpha
    self.beta = beta
    self.k = k
    

shangeth

class AlexNet(nn.Module):
  def __init__(self, classes=1000):
    super(AlexNet, self).__init__()
    self.features = nn.Sequential(
            nn.Conv2d(3, 64, kernel_size=11, stride=4, padding=2),
            nn.ReLU(inplace=True),
            nn.MaxPool2d(kernel_size=3, stride=2),
            nn.Conv2d(64, 192, kernel_size=5, padding=2),
            nn.ReLU(inplace=True),
            nn.MaxPool2d(kernel_size=3, stride=2),

shangeth

x = torch.randn(1, 3, 224, 224).uniform_(0, 1)
alexnet = AlexNet()
alexnet(x).size()

shangeth

from skimage import io
import matplotlib.pyplot as plt

image = io.imread('https://cdn3.bigcommerce.com/s-nadnq/product_images/uploaded_images/20.jpg')

plt.imshow(image)
plt.grid(False)
plt.axis('off')
plt.show()

shangeth

import random
from scipy import ndarray
import skimage as sk
from skimage import transform
from skimage import util

def random_rotation(image_array: ndarray):
    random_degree = random.uniform(-25, 25)
    return sk.transform.rotate(image_array, random_degree)

shangeth

def dropout(X, drop_probability):
    keep_probability = 1 - drop_probability
    mask = np.random.uniform(0, 1.0, X.shape) < keep_probability
    if keep_probability > 0.0:
        scale = (1/keep_probability)
    else:
        scale = 0.0
    return mask * X * scale
  
  

shangeth

import torch

x = np.random.randn(3,3)
x_tensor = torch.from_numpy(x)

dropout = torch.nn.Dropout(0.5)
dropout(x_tensor)

shangeth

X = np.random.rand(3,3)
print('Input X  = \n',X)

w = np.random.normal(loc=0.0, scale=0.01, size=(3,1))
print('\nInitialized weight w = \n',w)

bias = np.ones((3,1))
print('\nInitialized bias b = \n',bias)

z = np.dot(X, w) + bias z

shangeth

Deep Reinforcement learning

• Monte Carlo Prediction

  Monte Carlo Prediction method reinforcement learning agent for BlackJack Game.

• Temporal Difference methods

  Temporal Difference methods like

  • SARSA(0)
  • SARSAMAX(Q-Learning)

• Expected SARSA