Downloaded 10 times
Uploaded byShashikant Jagtap
CreateML Demo: Train Machine Learning Models
CreateML Talk at the London Test Automation in DevOps Meetup. https://www.meetup.com/londontad/events/258102828/
![Lect 1 Number systems and base conversions. [Autosaved].pptx](https://cdn.slidesharecdn.com/ss_thumbnails/lect1numbersystemsandbaseconversions-260111134109-67c2d865-thumbnail.jpg?width=640&height=640&fit=bounds)
CreateML Demo: Train Machine Learning Models
- 1. CreateML OR AITools for iOS App Testing? Shashikant Jagtap Twitter: Shashikant86 Blogs: https://shashikantjagtap.net
- 2. Topics •What is AIand ML •How AI-based tools test iOS Apps •Challenges in Testing iOS Apps with AI/ML •What’s is CreateML & How to Train ML Models with CreateML •CreateML - Demo
- 3. What is AIand ML Machine Learning Artificial Intelligence
- 4. Why QA NeedsAI/ML? DevOps QA with the pace of CI/CD is impossible Continuous DeliveryContinuous Deployment
- 5. How AI/ML HelpQA? AI will help existing QA Efforts Code-less Planning-less Flaky-less human-less? Infrastructure-less
- 6.
- 7. How AI/ML ToolsTest Web Apps? •Provide web app URL and test login credential •AI tool will crawl your web app and reports broken link, JS errors •Auto Generates Scenarios and Perform Journeys •Give Feedback from CI Server •No Code, No Infrastructure. Trainable Virtual QA Engineer
- 8. How AI/ML ToolsTest Mobile Apps? •Provide pre-built app binary (IPA, APK) •AI tool will install app on devices •Auto Generates Scenarios and Perform Journeys •Reports issues, crashes. Generate Screenshots & recordings •Give Feedback from CI Server •No Code, No Infrastructure.
- 9. iOS App Testing •UnitTesting ( XCTest) •Integration and Network Testing (XCTest, Fitnesse) •UITest (XCUITest, Appium, Calabash) •Non-functional testing •Heavy Manual Testing
- 10. Role of AIin iOS Testing •Reduce the burden on controlled testing •Free up time for exploratory testing •Great Coverage in limited time •Works well for faster feedback
- 11. Challenges •iOS Deployment andBuild Process •Apple Ecosystem •Training Models for iOS Apps •App Extensions
- 12. Welcome CreateML •Apple’s framework-Train ML Model •No need to learn TensorFlow or IBM Watson •Just few lines of Swift and data to train models
- 13. What You Need •maOSMojave •Xcode 10 •You are set!
- 14. How To TrainModels •Collect Training and Testing Data •Open Xcode Playground-macOS template •import libraries and start Training Models
- 15. How To TrainModels
- 16. Save Accurate Model •Onceachieved desired Accuracy •Save the model (.mlmodel) •Use this model for CoreML •No Need to Learn other ML tools
- 17. What Can BeTrained •Images •Tabular data (CSV, JSON) •Complex data if you are ML expert
- 18. Can CreateML TestiOS Apps •Images —> Visual Validation Testing •JSON -> API Testing of JSON Responses •Let’s Explore that into the Demo
- 19. It’s All AboutDATA https://www.evolvedmedia.com/tomato-tomato-how-do-you-say-data/
- 20.
- 21.
- 22.