Download as PDF, PPTX
Uploaded byIsabel Palomar
Creating a custom ML model for your application - DevFest Lima 2019
The document outlines the process of creating a custom machine learning model for applications, emphasizing deep learning concepts, data acquisition, model training, and evaluation techniques. It includes specifics on building an image classifier to recognize different beer types using transfer learning and various tools, including TensorFlow Lite for mobile applications. The author shares personal experiences and challenges in learning deep learning while providing practical resources and methods for model deployment.
![ANPARA THERMAL POWER STATION[1] sangam.pdf](https://cdn.slidesharecdn.com/ss_thumbnails/anparathermalpowerstation1sangam-251121115219-9261cde4-thumbnail.jpg?width=640&height=640&fit=bounds)
Creating a custom ML model for your application - DevFest Lima 2019
- 1. CREATING A CUSTOMMACHINE LEARNING MODEL FOR YOUR APPLICATIONS
- 2. Hello! I am IsabelPalomar 2
- 3. Agenda ● Challenges andInitial ideas ● Main Deep Learning concepts ● Using the model! ○ Android applications ○ In your iOS, web or backend applications … that’s all 3 ��
- 4. 1. CHALLENGES AND INITIALIDEAS Getting started with Machine Learning and Android
- 5. Production-ready for commonuse cases 5
- 6. I want tocreate a custom model HOW? 6
- 7. “ 77 @jeremyphoward @math_rachel I started aDeep Learning Course last year at the University of San Francisco. 💃 MY STORY: TAKING MY FIRST DEEP LEARNING CLASS
- 8. 8 🤯 After theclass….. The key outcome of this lesson is that we'll have trained an image classifier which can recognize pet breeds at state of the art accuracy. The key to this success is the use of transfer learning, which will be a key platform for much of this course. We also discuss how to set the most important hyper-parameter when training neural networks: the learning rate, using Leslie Smith's fantastic learning rate finder method. Finally, we'll look at the important but rarely discussed topic of labeling, and learn about some of the features that fastai provides for allowing you to easily add labels to your images. https://course.fast.ai/videos/?lesson=1
- 9. challenges…. ‐ Many courses,even basic, assume that you already know the subject. ‐ Reaching the final result without learning the basics is not good. 9
- 10. “When you arestarting to learn about Deep Learning it seems that there are thousands of concepts, mathematical functions and scientific articles that you have to read. 10 myths
- 11. 2. Main DEEP LEARNINGCONCEPTS Let’s understand the general concepts first
- 12. MACHINE LEARNING CONCEPTS DataTask Model 12 Loss Function Learning Algorithm Evaluation
- 13. Let’s take alook of the implementation We are going to build an app to classify the artisanal beers of Cervecería Colima Place your screenshot here 13
- 14. 1.- dATA Data isdistinct pieces of information which acts as a fuel 14
- 15. Formats Data can comein different formats depending on what you want to solve 15
- 16. DATA FOR OUREXAMPLE 16
- 17. How? Where dowe get data from? Data curation is the organization and integration of data collected from various sources. 17 Techniques You can use techniques like Questionnaires and surveys, conducting interviews, using data scraping and data crawling techniques.
- 18. Public datasets ● GoogleAI ● UCI ML Repository ● Data.gov.in ● Kaggle Where do we get data from? Crowdsourcing Marketplaces ● Amazon Mechanical Turk ● Dataturks ● Figure-eight 18
- 19. BACK TO OUREXAMPLE... ● Google Images ● https://github.com/hardikvasa/google-images-download ● https://forums.fast.ai/t/tips-for-building-large-image-datasets/26688 19
- 20. 2.- task The problemto resolve 20
- 21.
- 22. TASK FOR OUREXAMPLE 22 Classify Images of Artisanal Beers
- 23. Image classification A commonuse of machine learning is to identify what an image represents. The task of predicting what an image represents is called image classification. 23
- 24. 3.- model The mathematicalformulation of a task. 24
- 25. models 25 There are manymodels that are created over the years. Each model has its own advantages and disadvantages based on the type of data on which we are creating a model.
- 26. IMAGE CLASSIFICATION MODEL Animage classification model is trained to recognize various classes of images. 26 When we subsequently provide a new image as input to the model, it will output the probabilities of the image representing each of the types it was trained on.
- 27. An example outputmight be as follows: Beer type Probability Cayaco 0.02 Colimita 0.96 Piedra Lisa 0.01 Ticus 0.00 Paramo 0.01 27 Based on the output, we can see that the classification model has predicted that the image has a high probability of representing a Colimita Beer.
- 28. In this example,we will retrain a MobileNet. MobileNet is a a small efficient convolutional neural network. https://ai.googleblog.com/2017/06/mobilenets-open-source-models-for.html Model for our example 28
- 29. Retraining the mobileNetmodel 29 We use MobileNet model and retrain it. python3 -m scripts.retrain --bottleneck_dir=tf_files/bottlenecks --model_dir=tf_files/models/"${ARCHITECTURE}" --summaries_dir=tf_files/training_summaries/"${ARCHITECTURE}" --output_graph=tf_files/retrained_graph.pb --output_labels=tf_files/retrained_labels.txt --architecture="${ARCHITECTURE}" --image_dir=tf_files/beer_photos IMAGE_SIZE=224 ARCHITECTURE="mobilenet_0.50_${IMAGE_SIZE}" tHE RESULT...
- 30. USING THE RETRAINEDMODEL 3030 Evaluation time (1-image): 0.250s ticus (score=0.99956) paramo (score=0.00043) cayaco (score=0.00000) piedra lisa (score=0.00000) colimita (score=0.00000) python3 -m scripts.label_image --graph=tf_files/retrained_graph.pb --image=tf_files/beer_photos/ticus/"3. ticus.jpg"
- 31. 4.- loss function Howdo we know which model is better? Loss function (also known as the error) answers this question. 31
- 32. Classification losses: ● MeanSquare Error/L2 Loss ● Mean Absolute Error/L1 Loss Regression losses: ● Hinge Loss/Multi-class SVM Loss ● Cross Entropy ● Loss/Negative Log Likelihood LOSS FUNCTIONS To know which model is good for our data, we compute the loss function by comparing the predicted outputs to actual output. 32
- 33. 5.- learning algorithm TheLearning Algorithms also known as Optimization algorithms helps us to minimize Error 33
- 34. Is something youdo everyday... You are optimizing variables and basing your personal decisions all day long, most of the time without even recognizing the process consciously https://mitsloan.mit.edu/ideas-made-to-matter/how-to-use -algorithms-to-solve-everyday-problems 34
- 35. First Order Optimization Algorithms ●Gradient Descent Types of learning algorithms Second Order Optimization Algorithms ● Hessian https://towardsdatascience.com/types-of-optimization-algorithms-used-in-neural-networks-and- ways-to-optimize-gradient-95ae5d39529f 35
- 36. 6.- evaluation To computea score for our ML model, we need to evaluate it. 36
- 37. evaluation There are manyperformance metrics to achieve this. One simple such metric is Accuracy. 37
- 38. Evaluation/performance metrics 38 ● Confusionmatrix ● Area Under the ROC curve (AUC — ROC) ● Root Mean Squared Error (RMSE) (used in regression)
- 39. DEEP LEARNING CONCEPTS DataTask Model 39 Loss Function Learning Algorithm Evaluation
- 40. 3. USING THE MODEL Howto use our model in my apps
- 41. MACHINE LEARNING INYOUR APPS ● ML Kit For Firebase ● Core ML (Apple) ● TensorFlow Lite ● Cloud-based web services ● Your own service Place your screenshot here 41
- 42. TENSORFLOW LITE TensorFlow Liteis an open source deep learning framework for on-device inference. 42
- 43. USING THE RETRAINEDMODEL 4343 Evaluation time (1-image): 0.250s ticus (score=0.99956) paramo (score=0.00043) cayaco (score=0.00000) piedra lisa (score=0.00000) colimita (score=0.00000) python3 -m scripts.label_image --graph=tf_files/retrained_graph.pb --image=tf_files/beer_photos/ticus/"3. ticus.jpg"
- 44. TENSORFLOW LITE 44 TensorFlow Liteis a set of tools to help developers run TensorFlow models on mobile, embedded, and IoT devices. ● TensorFlow Lite converter ● TensorFlow Lite interpreter TensorFlow Lite converter Converts TensorFlow models into an efficient form for use by the interpreter
- 45. Command line: tflite_convert Startingfrom TensorFlow 1.9, the command-line tool tflite_convert is installed as part of the Python package. 45 pip install --upgrade "tensorflow==1.9.*"
- 46. Using TFLite Converter 46 Output:IMAGE_SIZE=224 tflite_convert --graph_def_file=tf_files/retrained_graph.pb --output_file=tf_files/optimized_graph.lite --input_format=TENSORFLOW_GRAPHDEF --output_format=TFLITE --input_shape=1,${IMAGE_SIZE},${IMAGE_SIZE},3 --input_array=input --output_array=final_result --inference_type=FLOAT --input_data_type=FLOAT
- 47.
- 48. 3.1 Use the model- android
- 49. Required files 49 Optimized graph cptf_files/optimized_graph.lite android/tflite/app/src/main/assets/graph.lite cp tf_files/retrained_labels.txt android/tflite/app/src/main/assets/labels.txt Labels
- 50. repositories { maven { url'https://google.bintray.com/tensorflow' } } dependencies { // ... compile 'org.tensorflow:tensorflow-lite:+' } TensorFlow Lite interpreter 50 android { aaptOptions { noCompress "tflite" noCompress "lite" } } The TensorFlow Lite interpreter is designed to be lean and fast. The interpreter uses a static graph ordering and a custom (less-dynamic) memory allocator to ensure minimal load, initialization, and execution latency. dependencies settings
- 51. Load model andcreate interpreter class ImageClassifier constructor(private val assetManager: AssetManager) { init { interpreter = Interpreter(loadModelFile(assetManager, MODEL_PATH)) labels = loadLabelList(assetManager) ... } } 51 // Name of the model file stored in Assets. const val MODEL_PATH = "graph.lite"; // Name of the label file stored in Assets. const val LABEL_PATH = "labels.txt";
- 52. cAMERA, Read thelabels….. 52 https://developer.android.com/training/camerax // Convert the image to bytes convertBitmapToByteBuffer(bitmap) // An array to hold inference results, to be feed into Tensorflow Lite as outputs. val recognitions = ArrayList<Result>() val recognitionsSize = Math.min(pq.size, MAX_RESULTS) for (i in 0 until recognitionsSize) recognitions.add(pq.poll()) return@flatMap Single.just(recognitions)
- 53. Show the results 53 //Get the results textToShow = String.format("n%s: %4.2f", label.key, label.value) // Label (In this case PARAMO) label.key // Value (In this case 1.0) label.value ticus (score=0.00000) paramo (score=1.00000) cayaco (score=0.00000) piedra lisa (score=0.00000) colimita (score=0.00000)
- 54. Call to action! Nowyou are ready to start building your first custom ML model. 54
- 55. 3.1 Use the model- other applications iOS, web or backend applications
- 56. SOMETHING IS HAPPENING….. 56 TheAI industry is constantly changing and evolving
- 57. Tensorflow lite options- Mobile & iot TensorFlow Lite provides all the tools you need to convert and run TensorFlow models on mobile, embedded, and IoT devices. 57
- 58. Tensorflow platform 58 TensorFlow isan end-to-end open source platform for machine learning. ● Web Apps ● Back end - APIs https://www.tensorflow.org/
- 59. 59 YOU CAN BUILDGREAT THINGS! https://ai.google/social-good/ There a lot of problems to solve! 59
- 60.
- 61. 61 THANKS! You can findme at @isabelpalomar isabelpalomarb@gmail.com