Blogs

Keypoints What is Convolution? Convolution is one of the main building blocks of a CNN. The term convolution refers to the mathematical combination of two functions to produce a third function. It merges two sets of information. In the case of a CNN, the convolution is performed on the input data with the use of a “filter” or “kernel”(these terms are used interchangeably) to then produce a “feature map”. What are Convolutional Neural Networks (CNN)?...

Keypoints Introduction Recommendation systems are one of the predominant systems in market like Netflix, amazon and Walmart. And also this applies to offline systems such as which product goes in which row to capture the users. And it is one of the challenging problems. The solution for that problem is called Collaborative Filtering. Collaborative Filtering The Collaborative Filtering technique refers to looking at what products the current user has used or liked, find other users that have used or liked similar products, and then recommend other products that those users have used or liked....

Keypoints Introduction It is better to fail fast than very late. And it is always better to run more experiments on a smaller dataset rather running a single experiment on a large dataset. This chapter introduces a new dataset called Imagenette1 Imagenette is a subset of the original Imagenet dataset but has only 10 categories of classes which are very different. This dataset has full-size, full-color images, which are photos of objects of different sizes, in different orientations, in different lighting, and so forth....

Keypoints Multi-label classification Multi-label classification refers to the problem of identifying the categories of objects in images that may not contain exactly one type of object. There may be more than one kind of object, or there may be no objects at all in the classes that you are looking for. See two examples below where we have a bear dataset with a dog included named bear and another example where the cat is classified as cat and horse....

Keypoints The Keytopics in the blogpost: Presizing / Augmentation of Images Datablock Cross-Entropy Loss Presizing / Augmentation of Images The main idea behind augmenting the images is to reduce the number of computations and lossy operations. This also results in more efficient processing on the GPU. To make the above possible we need our images to have same dimensions, so they can be easily collated. Some of the challenges in doing the augmentation is that when we resize, the data could be degraded, new empty zones are introduced etc....