Deep Learning Course with Flutter & Python - Build 6 AI Apps

Use Teachable Machine by Google to build a neural network that detects yes or no in sign language from webcam data, balance the dataset, and tune epochs and learning rate.

Develop a cat-vs-dog detector using a convolutional neural network built with Teachable Machine, featuring drag-and-drop data and one-click training for rapid AI prototyping.

Explore building an on-device cat and dog classifier in flutter by using Teachable Machine to train a neural network, export a tflite model, and integrate image picker and splash screen.

Configure a Flutter project by cleaning default files, creating an assets folder and pubspec.yaml entries, then implement a splash screen with a title, image, and navigation to home page.

Fix the min sdk error by updating android/app/build.gradle to minSdkVersion 19, add the flight package and image_picker dependency, and begin implementing flight functions in home.dart.

Set up a flutter ui scaffold with a padded container and column layout, presenting a teachable machine convolutional neural network and an image to detect dogs and cats.

In this Flutter and Python deep learning lecture, add two buttons to pick an image from gallery or take a photo to feed into the neural network, with state updates.

Capture a photo from your device, store it in a variable, input the image into a neural network, and train the model on teachable machines.

Acquire a cats-and-dogs image dataset from Kaggle and train a neural network in Teachable Machine with images. Export the model to TensorFlow for use in Flutter and Python apps.

Load the model in a Flutter app, then classify images using a tf lite model and update a loading state while preventing memory leaks with proper disposal.

Pass an image from the gallery to the neural network, call the classify image function, manage loading state, and display the predicted outputs in the app UI.

Finish and test your flutter app by adding a loading state to show the selected image and the neural network's prediction label.

Develop a flower type detector app by customizing a neural network in Colab, using Python in your browser to build and run the model without local installations.

Set up the project by duplicating the initial app, renaming it to flower detector, importing assets, and building a gradient splash screen in Flutter to host a neural network.

Set up the home screen in a Flutter app with a linear gradient background and a centered column for text and image preview. Shows a convolutional neural network.

Create a Flutter home screen that displays a model's prediction label when output is available, using size boxes and text styling, while preparing reusable buttons for the next steps.

Develop a responsive Flutter home screen with a styled take a photo button, using media query, gesture detector, and padding; next, build neural networks in Google Colab.

Learn how to use Google Colab to build neural networks, run Python code, install packages with pip, set GPU runtime, import files, and work with Jupyter notebooks.

Preprocesses image data for a neural network by downloading a flower photos dataset, resizing images to 224, creating train/validation generators, and generating labels for transfer learning with MobileNet.

Learn how to build and export a transfer learned MobileNet model for image classification, using 224×224 RGB inputs, five classes, and tf lite conversion.

Finish and test the flutter and tf lite app by updating the tf lite model and labels, rebuilding, and validating five flower classes with expansion to 131 fruits and vegetables.

Build a convolutional neural network to detect up to 131 fruits and vegetables using transfer learning, expanding from five categories, and watch a live demo classifying a red delicious apple.

Copy and paste the previous app to create a new neural network project named Ifrit. Then update colors, texts, and assets in Flutter to match the new network.

Set up the home screen by updating titles, assets, and a blue gradient theme, run the app, and prepare a 131 fruit and vegetable classifier.

Download the fruits 360 dataset, unzip and organize training and test folders, create a python scripts directory, and start a local TensorFlow workflow in Jupyter to build the neural network.

Set up the dataset path, image size, and batch; augment data with rotation, shifts, and flips, then create the training generator and save the labels before building the neural network.

Build a neural network with transfer learning using MobileNet v2, freeze the base, add top layers for 131-class softmax classification, and train with Adam, categorical cross-entropy, and accuracy.

Explore sentiment analysis by sending phrases to an API and retrieving emotion predictions, illustrated with a demo that shows joy and discusses applications like stock market insights from tweets.

Create a new flutter project for sentiment analysis, add a splash screen dependency, configure main.dart, integrate assets and a gradient background, and prepare the home page.

Learn to build the home screen in Flutter by setting up a scaffold, gradient container, and a padded column with responsive sizing and a sentiment analysis title.

Build the home screen with a loading boolean and a text field connected to a text editing controller. Send input to an API for sentiment analysis and display the results.

Add a button using a gesture detector that sends the textfield input to a sentiment analysis API and displays the result on screen, with width set via media query.

Set up an api service in a flutter app using rapid api’s text analysis, obtain your api key, configure base url and headers, and post requests with json parsing.

Integrate an emotion analysis API into a Flutter app by configuring an API service, sending user text, and displaying predicted emotions such as joy and sadness.

Set up a Flutter project, add dependencies, replace starter code, implement a splash screen and home page, and prepare assets to enable live captions.

Set up the home screen with a scaffold, a linear gradient background, and a centered column inside a container for text generator and image to text generator.

Set up a Flutter home screen with an image asset and caption generator buttons: live caption generator, take a photo caption generator, and take a photo from gallery caption generator.

Develop and wire the home screen in a Flutter and Python deep learning app, including image picking, loading state management, and parsing JSON from an HTTP API to update results.

Build an API on Red Hat OpenShift that accepts an image via post and returns a JSON string with the text inside the image.

Builds an asynchronous function in Flutter to fetch a response by uploading an image via a multipart post request to an API, including mime type checks.

Toggle the loading state to switch containers, display an image with border radius, and show fetch response text; enable image picking from camera or gallery for a live camera inference.

Set up live camera functionality in flutter by adding the camera dependency, detecting cameras, initializing a back camera with medium resolution, and capturing photos every five seconds for http requests.

Finalize the live camera feature in a Flutter app with a scaffolded gradient UI, a live preview, periodic predictions, and navigation to the live captions screen.

Set up a Flutter project for a sketch-driven image generator using a generative adversarial network, add dependencies (http, file picker, path provider, cached network image), and scaffold the home screen.

Set up the home screen scaffold in a Flutter app using a stack with a gradient container and a centered 256x256 drawing canvas, controlled by a loading flag.

Build a drawing canvas in Flutter by creating a drawing area and a custom painter that renders a list of points as lines to enable sketching.

Add drawing with a gesture detector tracking onPanDown, onPanUpdate, onPanEnd, updating points. Render the sketch with a custom painter using black stroke and round caps.

Implement a save-to-image workflow for a user sketch by rendering points on a canvas, converting the drawing to PNG bytes, and saving it locally for neural network input.

Add an icon button to clear the sketch by clearing points with set state, enabling draw and clear functionality, and prepare to show the saved image on the next screen.

Manage loading state to display a generated image in Flutter, via a back button, image bytes, memory image, and an http request to a Flask neural network server.

Train and export the neural network after 2320 steps; more data improves performance, load samples, and test predictions to compare generated and expected images after one epoch.

learn how to build a basic flask server to deploy a neural network, including a hello world route and a rest api for flutter applications.

Build a Flask server that loads a saved neural network and preprocesses images, exposing generate and download routes to predict and serve outputs for a Flutter app.

Extend a Flutter app to call a Flask API, test the server, upload an image via multipart request to port 5000, decode the JSON response, and display the generated image.

Set up a flask server for a flutter app, configure host ip, handle image preprocessing and input shape, and debug model deployment issues end-to-end.

Add a file input button to feed a selected image into your neural network and generate a realistic output, demonstrated with a shoe sketch and a file picker integration.

Explore edges2shoes and other pix2pix datasets from Berkeley, including cityscapes, shoes, and handbags, and export and organize training, validation, and test folders for a pix2pix neural network using TensorFlow.