Words About Dev

In machine learning, transfer learning is a technique where a pre-trained model, developed for one task, is reused as the starting point for another similar task. This can be a huge timesaver, as it reduces the time and computational resources required to train models from scratch. In this blog post, we will discuss how to use this technique to improve performance in image classification tasks, specifically using TensorFlowJS and the MobileNetV2 model....

If you’re building an Android app that requires a map, chances are you’ll need to display it in a Fragment at some point. Fortunately, adding a MapView to a Fragment is relatively easy to do with the Google Maps SDK for Android. In this post, we’ll walk you through the steps to add a MapView to a Fragment in your Android app. Create a new Fragment in Android Studio To get started, you’ll need to create a new Fragment in your Android Studio project....

When developing Android applications that open URLs, it is essential to check if a web browser that supports the ACTION_VIEW intent is available on the device. Failing to do so can result in an ActivityNotFoundException or a NullPointerException that may crash your application. This is especially true if the targeted device is under Mobile Device Management (MDM) control, as the MDM administrator may not have installed a browser on the device....

As a developer working with REST APIs, you have likely encountered situations where you need to deserialize the response data from a web server. In Kotlin, the most common approach to deserialize response data is by defining a model class that mirrors the server’s response. While it may be tempting to declare all the properties that the server returns in your Kotlin model, it is not always the best approach. Yesterday, a customer reported to me that our app was crashing during login....

Deep learning models have been successful in a wide range of applications, from image classification to natural language processing. However, training these models from scratch can require a large amount of labeled data and computational resources. Fortunately, there are many pre-trained deep learning models available that can be used for a variety of tasks, and the tf.keras.applications module provides an easy way to load these models in TensorFlow. In this blog post, we’ll look at a specific implementation of a function that loads pre-trained deep learning models using the tf....