내 손 위의 딥러닝_iOS에 딥러닝 심기
•
2 likes•974 views
2018년 6월 28일 첫번째 함께하는 딥러닝 컨퍼런스에서 발표한 "내 손 위의 딥러닝_iOS에 딥러닝 심기"입니다. 앞으로 점점 더 가속화 될 모바일 딥러닝에 대해 소개합니다. 주요 내용은 아래와 같습니다. 1. 모바일 딥러닝이란 무엇인가 2. 모바일 딥러닝의 장점 3. 모바일 딥러닝의 등장 배경 4. 모바일 딥러닝 활용 방안 5. 모바일 딥러닝 iOS 구축 사례 6. 모바일 딥러닝 한계 및 극복 방법 궁금한 점이 있다면 github 혹은 메일을 활용해주세요.
Recommended
More Related Content
Similar to 내 손 위의 딥러닝_iOS에 딥러닝 심기
Similar to 내 손 위의 딥러닝_iOS에 딥러닝 심기 (20)
More from Mijeong Jeon
Recently uploaded
Recently uploaded (20)
내 손 위의 딥러닝_iOS에 딥러닝 심기
- 2. 1ST
- 3. 1ST
- 4. 1ST
- 5. 1ST
- 6. 1ST
- 8. 1ST
- 11. 1ST 0.1 s 😎 1 s 5 s 😫 😯
- 12. 1ST Geekbench 4 - cross-platform processor benchmark
- 13. 1ST
- 15. 1ST Classification Object Detection Regression Decision Tree Classifier Logistic Regression Random Forest Regression Linear Regression Random Forest Classifier SVM
- 17. 1ST AgeNet Input: Image Output: String 43.5 MB https://github.com/cocoa-ai/FacesVisionDemo
- 18. 1ST DocumentClassification Input: String Output: String 1.4 MB https://github.com/toddkramer/DocumentClassifier
- 19. 1ST Exermote Input: Accelerations Output: Array 80 KB https://github.com/Lausbert/Exermote/tree/master/ExermoteInference
- 20. 1ST Food101 Input: Image Output: String 83.3 MB https://github.com/ph1ps/Food101-CoreML
- 21. 1ST RN1015k500 Input: Image Output: Dictionary 284.4 MB https://github.com/awslabs/MXNet2CoreML_iOS_sample_app
- 22. 1ST
- 23. 1ST
- 24. 1ST
- 26. 1ST “Don’t be a hero!” -Andrej Karpathy
- 27. 1ST
- 28. 1ST
- 29. 1ST
- 30. 1ST
- 31. 1ST
- 35. 1ST
- 36. 1ST Xcode Swift
- 37. 1ST
- 39. 1ST
- 40. 1ST
- 41. 1ST
- 42. 1ST from keras.preprocessing.image import ImageDataGenerator # 1. train_datagen = ImageDataGenerator(rescale=1./255) train_generator = train_datagen.flow_from_directory( ‘data_set/train’, // target_size=(28, 28), // batch_size=10, // class_mode=‘categorical') // : https://tykimos.github.io/
- 43. 1ST from keras.preprocessing.image import ImageDataGenerator # 1. train_datagen = ImageDataGenerator(rescale=1./255, rotation_range=10, width_shift_range=0.2, height_shift_range=0.2, shear_range=0.7, zoom_range=[0.9, 2.2], horizontal_flip=True, vertical_flip=True, fill_mode='nearest') : https://tykimos.github.io/
- 44. 1ST # 2. from keras.models import Sequential from keras.layers import Dense, Flatten from keras.layers.convolutional import Conv2D, MaxPooling2D model = Sequential() // model.add(Conv2D(32, // kernel_size=(3, 3), // activation=‘relu’, // input_shape=(28, 28, 3)) // ( , ) model.add(Conv2D(64, (3, 3), activation=‘relu')) // model.add(MaxPooling2D(pool_size=(2, 2))) // model.add(Flatten()) // -> model.add(Dense(128, activation=‘relu’)) // model.add(Dense(3, activation=‘softmax')) // : https://tykimos.github.io/
- 45. 1ST # 3. model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=[‘accuracy']) // # 4. model.fit_generator( train_generator, // steps_per_epoch=60, // / (= 600/10) epochs=150) // # 5. score=model.evaluate_generator(test_generator, steps=3) : https://tykimos.github.io/
- 46. 1ST
- 47. 1ST # 6. coreml import coremltools coreml_model = coremltools.converters.keras.convert(model, input_names=‘image’, // = output_names=‘class', image_input_names = ‘image', class_labels = ['circle', 'rectangle', ‘triangle'], // is_bgr=True) coreml_model.save(‘shape_detect_with_keras.mlmodel') //
- 48. 1ST
- 49. 1ST
- 50. 1ST # 7. coreml import CoreML let model = shape_detect_with_keras() guard let output: shape_detect_with_kerasOutput = try? model.prediction(image: image) else { return } https://github.com/MijeongJeon
- 51. 1ST
- 52. 1ST
- 53. 1ST
- 57. 1ST Apple
- 58. 1ST Apple
- 59. 1ST Apple Image Classification Text Meaning Find Relationship
- 60. 1ST Apple
- 62. 1ST IBM
- 63. 1ST IBM
- 64. 1ST Microsoft
- 65. 1ST Microsoft
- 66. 1ST
- 67. 1ST class func compileModel(at: URL) -> URL convert_neural_network_spec_weights_to_fp16(model_spec)
- 68. 1ST
- 69. 1ST IBM
- 71. 1ST
- 72. 1ST
- 73. 1ST 🤓