Forward Single Image with airSlate SignNow

Forward single image

welcome to this series on network programming with hi George in this episode we will see how we can use our convolutional neural network to generate an output prediction tensor from a sample input of our data set without further ado let's get started at this point in the series we've finished building a model and technically we could jump right into the training process from here however let's work to better understand how our network is working right out of the box and then once we understand our network a little more deeply we'll be better prepared to understand the training process the first step is to understand Ford propagation for propagation is the process of transforming an input tensor to an output tensor at its core a neural network is a function that map's an input tensor to an output tensor and for propagation is just a special name for the process passing an input tensor to the network and receiving the output from the network as we've seen neural networks operate on data in the form of tensors the word Ford propagate is used to indicate that the input tensor data is transmitted through the network in the four direction for our network what this means is simply passing our input tensor to the network and then receiving the output tensor to do this we pass our sample data to the networks or method this is why the form method has the name Ford the execution of the form method is the process of Ford propagation if you're following the series we know by now that we don't call the form method directly rather we call the network instance instead make sure to see the episode on callable PI torch neural networks for more details on this let's discuss the word propagate the word propagate means to Moo or transmitted through some medium in the case of neural networks data propagates through the layers of the network there is a notion of backward propagation or back prop as well which makes the term for propagation suitable as a first step during the training process back propagation occurs after for propagation for propagation is the process of passing an input image tensor to the four method that we implemented in the last episode the output we receive from the four method is the network's prediction in the episode on data sets and data loaders we saw how to access a single sample image tensor from our training set and more importantly how to access a batch of image tensors from our data loader now that we have our network defined and our four method implemented we're ready to pass an image to our network to get a prediction let's get our code set up we need our imports our training set and our network [Music] [Music] [Music] [Music] [Music] [Music] [Music] all right before we begin we are going to turn off PI torches a gradient calculation feature...