Why are deep neural networks hard to train (Vanishing & Exploding gradient )As we keep increasing the depth of the network with more and more hidden layers, we are increasing the network parameters and the network…May 27, 2020May 27, 2020
Cross Entropy (What made use look for a new cost function)How did we find cross entropy ?May 20, 2020May 20, 2020
What is Activation FunctionWe know activation functions like sigmoid, softmax, relu… , why do we use these activation functions. Why can’t we simply pass the…Apr 26, 2020Apr 26, 2020
Back-propagation Part 2In part 1 we understood how moving in the direction of gradient of a function can be used to find the max value of that function.Apr 25, 2020Apr 25, 2020
Gradient Descent Part 1 (What is Gradient)Gradient Descent is a strategy used in Neurol Networks to minimize the cost function. When I first heard about this work Gradient Descent…Apr 16, 2020Apr 16, 2020
Published inAnalytics VidhyaUse Pytorch to create an image captioning model with CNN and seq2seq LSTM on Google Colab GPUThe COCO dataset is used. We used the year 2014 data.Dec 20, 2019Dec 20, 2019
Simple Client Based Custom Object Detection APISteps to setup and use the Custom Object Detection APIOct 21, 2019Oct 21, 2019
