Swift Lille - CoreML / CreateML

Transcript

1. Image classification in real world Car rental damage report
8. // Load the ML model through its generated class guard let model = try? VNCoreMLModel(for: CarClassifier().model) else { fatalError("can't load Places ML model") } // Create a Vision request with completion handler let request = VNCoreMLRequest(model: model) { [weak self] request, error in guard let results = request.results as? [VNClassificationObservation] else { return } let topResults = results[0..<5] // Update UI on main queue DispatchQueue.main.async { [weak self] in self?.answerLabel.text = topResults .map { "\(Int($0.confidence * 100))% \($0.identifier)" } .joined(separator: "\n") } } // Run the Core ML classifier on global dispatch queue let handler = VNImageRequestHandler(ciImage: image) DispatchQueue.global(qos: .userInteractive).async { do { try handler.perform([request]) } catch { print(error) } }
9. // Load the ML model through its generated class guard let model = try? VNCoreMLModel(for: CarClassifier().model) else { fatalError("can't load Places ML model") } // Create a Vision request with completion handler let request = VNCoreMLRequest(model: model) { [weak self] request, error in guard let results = request.results as? [VNClassificationObservation] else { return } let topResults = results[0..<5] // Update UI on main queue DispatchQueue.main.async { [weak self] in self?.answerLabel.text = topResults .map { "\(Int($0.confidence * 100))% \($0.identifier)" } .joined(separator: "\n") } } // Run the Core ML classifier on global dispatch queue let handler = VNImageRequestHandler(ciImage: image) DispatchQueue.global(qos: .userInteractive).async { do { try handler.perform([request]) } catch { print(error) } }
10. // Load the ML model through its generated class guard let model = try? VNCoreMLModel(for: CarClassifier().model) else { fatalError("can't load Places ML model") } // Create a Vision request with completion handler let request = VNCoreMLRequest(model: model) { [weak self] request, error in guard let results = request.results as? [VNClassificationObservation] else { return } let topResults = results[0..<5] // Update UI on main queue DispatchQueue.main.async { [weak self] in self?.answerLabel.text = topResults .map { "\(Int($0.confidence * 100))% \($0.identifier)" } .joined(separator: "\n") } } // Run the Core ML classifier on global dispatch queue let handler = VNImageRequestHandler(ciImage: image) DispatchQueue.global(qos: .userInteractive).async { do { try handler.perform([request]) } catch { print(error) } }
11. // Load the ML model through its generated class guard let model = try? VNCoreMLModel(for: CarClassifier().model) else { fatalError("can't load Places ML model") } // Create a Vision request with completion handler let request = VNCoreMLRequest(model: model) { [weak self] request, error in guard let results = request.results as? [VNClassificationObservation] else { return } let topResults = results[0..<5] // Update UI on main queue DispatchQueue.main.async { [weak self] in self?.answerLabel.text = topResults .map { "\(Int($0.confidence * 100))% \($0.identifier)" } .joined(separator: "\n") } } // Run the Core ML classifier on global dispatch queue let handler = VNImageRequestHandler(ciImage: image) DispatchQueue.global(qos: .userInteractive).async { do { try handler.perform([request]) } catch { print(error) } }
12. // Load the ML model through its generated class guard let model = try? VNCoreMLModel(for: CarClassifier().model) else { fatalError("can't load Places ML model") } // Create a Vision request with completion handler let request = VNCoreMLRequest(model: model) { [weak self] request, error in guard let results = request.results as? [VNClassificationObservation] else { return } let topResults = results[0..<5] // Update UI on main queue DispatchQueue.main.async { [weak self] in self?.answerLabel.text = topResults .map { "\(Int($0.confidence * 100))% \($0.identifier)" } .joined(separator: "\n") } } // Run the Core ML classifier on global dispatch queue let handler = VNImageRequestHandler(ciImage: image) DispatchQueue.global(qos: .userInteractive).async { do { try handler.perform([request]) } catch { print(error) } }
13. // Load the ML model through its generated class guard let model = try? VNCoreMLModel(for: CarClassifier().model) else { fatalError("can't load Places ML model") } // Create a Vision request with completion handler let request = VNCoreMLRequest(model: model) { [weak self] request, error in guard let results = request.results as? [VNClassificationObservation] else { return } let topResults = results[0..<5] // Update UI on main queue DispatchQueue.main.async { [weak self] in self?.answerLabel.text = topResults .map { "\(Int($0.confidence * 100))% \($0.identifier)" } .joined(separator: "\n") } } // Run the Core ML classifier on global dispatch queue let handler = VNImageRequestHandler(ciImage: image) DispatchQueue.global(qos: .userInteractive).async { do { try handler.perform([request]) } catch { print(error) } }
14. // Load the ML model through its generated class guard let model = try? VNCoreMLModel(for: CarClassifier().model) else { fatalError("can't load Places ML model") } // Create a Vision request with completion handler let request = VNCoreMLRequest(model: model) { [weak self] request, error in guard let results = request.results as? [VNClassificationObservation] else { return } let topResults = results[0..<5] // Update UI on main queue DispatchQueue.main.async { [weak self] in self?.answerLabel.text = topResults .map { "\(Int($0.confidence * 100))% \($0.identifier)" } .joined(separator: "\n") } } // Run the Core ML classifier on global dispatch queue let handler = VNImageRequestHandler(ciImage: image) DispatchQueue.global(qos: .userInteractive).async { do { try handler.perform([request]) } catch { print(error) } }
19. TOP 7 Amazing Tips About Machine Learning Pour les nuls (aka iPhone developers)
21. Image classification Dossiers Apprentissage Nouvelle image Classes
23. Chat Table Fenêtre Lampe de bureau
24. Chat Table Fenêtre Lampe de bureau Aile avant Jante arrière Pneu
25. #1 le Machine Learning n’est pas réservé aux data-scientists
28. Mon Dataset 41 269 images 62,63 Go La même voiture Les mêmes couleurs La même luminosité (très sombre) Déjà classifié par les utilisateurs
31. #2 Vous avez un dataset Ou vous devriez
32. UX proposition Est-ce une jante ? Oui bien sûr ! Non pas du tout
33. Create ML Top 1 - From 1 34 %
34. Jante Pneu Aile Trappe essence
35. Create ML Top 1 - From 1 34 % Top 1 - From 5 91 %
36. UX proposition v2 Something else Jante arrière Jante avant Pneu avant
37. Jante Toit Aile Trappe essence
38. Create ML Top 1 - From 1 34 % Top 1 - From 5 91 % Top n - From n 78 %
39. #3 Ne faites pas confiances Utilisez les résultats comme des indications
43. Base Merge Classes Top 1 - From 1 34 % 37 % Top 1 - From 5 91 % 93 % Top n - From n 78 % 81 %
44. #4 Merger les classes Si vous ne pouvez pas le voir, le ML non plus
45. Partner app
46. Base Partner data Top 1 - From 1 37 % 34 % Top 1 - From 5 93 % 89 % Top n - From n 81 % 75 %
47. Base Partner data Partner corrected data Top 1 - From 1 37 % 34 % 41 % Top 1 - From 5 93 % 89 % 94 % Top n - From n 81 % 75 % 84 %
48. #5 Netoyez vos données La taille ne compte pas toujours
50. https://medium.com/@mhausherr_53256/le-machine-learning-sur-mobile-au-service-de-la-location-de-voiture-6f8176e2e285
51. Inception v3 - MobileNet Inception v3 MobileNet Create ML Top 1 - From 5 91 % 90 % 91 % Top n - From n 79 % 18 % 78 % Size 87 Mo 2.2 Mo 1.5 Mo Training time 2h41 2h38 1h50
52. #6 Utilisez Create ML C’est top
53. #1 Le ML c’est aussi pour vous #2 Vous avez déjà un dataset #3 Ne faites pas confiance aux résultats #4 Si vous ne le voyez pas, le ML non plus #5 Nettoyez vos données #6 Utilisez Create ML #7 Faites attention sur la route
54. #1 Le ML c’est aussi pour vous #2 Vous avez déjà un dataset #3 Ne faites pas confiance aux résultats #4 Si vous ne le voyez pas, le ML non plus #5 Nettoyez vos données #6 Utilisez Create ML #7 Faites attention sur la route
55. Mathieu Hausherr @mhausherr