Visão Geral

Keras é uCurso Keras Foundations & Advanced. ma estrutura de programação popular de alto nível para aprendizado profundo que simplifica o processo de construção de aplicativos de aprendizado profundo. Em vez de codificar no TensorFlow de baixo nível e fornecer todos os detalhes, Keras fornece um wrapper de interface de programação simplificado sobre o Tensorflow. No curso de treinamento Deep Learning with Keras , você aprenderá a usar Keras para aprendizado profundo e aprendizado de máquina, CNN, RNN, e explorará muitos modelos poderosos de aprendizado profundo pré-treinados incluídos no Keras.

No Keras - curso de treinamento avançado, você aprenderá a API funcional Keras para construir modelos complexos, ajuste de hiperparâmetros com Keras e retornos de chamada automáticos de Keras.

No curso de treinamento Keras - Modelagem de Dados Sequenciais , você aprenderá análise de dados sequenciais usando Keras.

No curso de treinamento Keras - Deep Learning for Image Classification , você aprenderá a construir modelos de classificação de imagens básicos a avançados usando Keras.

Objetivo

Ao participar do Curso Keras Foundations & Advanced, os participantes aprenderão:

• Rede Neural com Keras
• Construa um modelo de regressão preditiva com Keras
• Construa um modelo de classificação com Keras
• Construa um modelo de classificação de imagens CNN com Keras
• Transferir aprendizagem com Keras
• Construa um modelo de classificação RNN com Keras
• Ajuste de hiperparâmetros com Keras e Auto-Keras
• Retornos de chamada
• API funcional para construir modelos complexos
• Backend Keras para invocar operações Tensorflow
• Auto-Keras
• Recapitulação de RNN e LSTM
• Convolução 1D
• Modelo de sequência de sequência 2
• Mecanismo de Atenção
• Rede Neural com Keras
• Construa um modelo de classificação de imagens CNN com Keras
• API funcional
• Transferir aprendizagem com Keras

Publico Alvo

• Cientistas de Dados
• Analistas de dados
• Engenheiros

Pre-Requisitos

• Python básico
• Aprendizado de máquina

Materiais

Conteúdo Programatico

Getting Started on Keras

1. What is Keras
2. Keras vs TensorFlow
3. Install and Run Keras

Predictive Model with Neural Network

1. What is Neural Network (NN)?
2. Scaling Data for Regression
3. Loss Function and Optimizer
4. Regression Predictive Model with NN

Classification Model with Neural Network

1. One Hot Encoding
2. Cross Entropy and SoftMax
3. Classification Model with NN

Convolutional Neural Network (CNN)

1. What is CNN?
2. ImageDataGenerator
3. Image Classification Model with CNN
4. Data Augmentation and Dropout

Transfer Learning

1. What is Transfer Learning
2. Keras Pre-Trained Models
3. Fine Tuning Pre-Trained Models

Recurrent Neural Network (RNN)

1. What is RNN?
2. Long Term Dependencies
3. LSTM and GRU Cells
4. RNN Classification Model on IMDB datasets

Recap on Keras Basic

1. Sequential Model
2. Feedforward Neural Network (NN)
3. Convolutional Neural Network (CNN)
4. Recurrent Neural Network (RNN)

Functional API

1. What is Functional API
2. Code Sequential Models with Functional API
3. Fine Tune Transfer Learning with Functional API
4. Implement GAN with Functional API
5. Create Multi Input and Output Model

Callbacks

1. Keras Callbacks
2. ModelCheckPoint Callback
3. Tensorboard Callback

Data Generator

1. Image Class Generator
2. Fit Generator
3. Flow from Directory Generator
4. Custom Data Generator

Keras Backend

1. What is Keras Backend
2. Keras Backend Commands
3. Create Custom Loss Function

Word Embedding

1. One Hot Encoding of Words
2. Word Embedding
3. Pre-Trained Word Embedding

RNN and LSTM

1. Recurrent Neural Network (RNN)
2. Long Short Term Memory (LSTM) and GRU
3. Stacked RNN
4. Bidirectional RNN
5. Case Studies on Time Series Prediction with LSTM

1D Convolution

1. 1D Convolution on Sequential Data
2. Combining 1D Convolution and RNN

Sequence To Sequence Model

1. What is Seq2Seq Model
2. Attention Mechanism

Keras Basics

1. What is Keras
2. Keras vs TensorFlow
3. Google Colab
4. Install and Run Keras on Google Colab

Image Classification Model with Feedforward Neural Network (NN)

1. What is Feedforward NN
2. One Hot Encoding
3. Cross Entropy and SoftMax
4. MNIST Dataset
5. NN Image Classification NN Model for HandWritten Digits

Image Classification with Convolutional Neural Network (CNN)

1. What is CNN?
2. CNN Architecture
3. CNN Image Classification for HandWritten Digits
4. Image Class Generator and Fit Generator
5. CNN Image Classification for Cats and Dogs Images
6. Solving Overfitting with Dropout & Data Augmentation
7. Mini Project on Image Classification

Image Classification with Transfer Learning

1. What is Transfer Learning
2. Image Classification with Pre-Trained Models
3. Fine Tune Pre-Trained Models
4. Mini Project on Transfer Learning

Keras Functional CNN Model

1. What is Functional API
2. Split CNN Model for Image Classification
3. Mini Project on Functional CNN Model

Object Detection with Mask R-CNN

1. Overview of R-CNN Models
2. Mask R-CNN Demo