End to End Deep Learning for Computer Vision

Description
Content
Speaker
COURSE OBJECTIVES:
As applied engineering just knows building models are not sufficient for production grade software, these roles focus on core principles, best practices, design patterns, and expertise with a framework and toolset, such as deploy models, and scale for your fast growing applications/services.
This two day immersive instructor-led training will teach everything you need to know to become a software engineer in computer vision and deep learning! You will learn:
  • Recognize problems can be solved with deep learning and Select right technique for problems
  • Select the right technique for the problems.
  • Master deep learning algorithms, models and computer vision tech
  • Master the most popular tools like numpy, Keras, Tensorflow, and openCV
  • Master google cloud machine learning pipelines
  • This training is packed with practical exercises and code labs. not only will you learn theory, but also get hands-on practice building your own models, tuning models, and serving models

    COURSE SCHEDULE:
    • June 15th Sat 9am-4:30pm PT
    • June 16th Sun 9am-4:30pm PT

    COURSE INCLUDE:
    • Practical walkthroughs that present solutions to actual, real-world image classification problems and challenges
    • Hands-on tutorials (with lots of code) that not only show you the algorithms behind deep learning for computer vision but their implementations as well
    • A no-nonsense teaching style that is guaranteed to cut through all the cruft and help you master deep learning for image understanding and visual recognition
    • End to end machine learning pipeline from building models to deploy and serve models
    • light breakfast, lunch and coffee breaks

    COURSE CONTENT:
    Check the content tab for full course outlines.

    WHO SHOULD LEARN:
    Developers interested in Computer Vision and Deep Learning

    PREREQUISITE:
    Python

    Day 1 (9am -4:30pm)
    Module 1: Computer Vision Models
    • Neural Networks
      • Activation, loss function
      • classifier
      • Flattening, overfitting, dropout
      • code lab with Keras
    • Convolutional Neural Networks
      • resize, feature detection
      • filters, strides, pooling
      • VGG, ResNet
      • Batch normalization
      • code lab with Keras
    • Wide Convolutional Neural Networks
      • inception, ResNeXt
      • code lab
    • Advanced CNNs
      • pre-stems, DenseNet, MobileNet
      • code lab
    Module 2: Computer Vision Data Engineering
    • Data Collection & Assembly
      • best practice
      • unbalanced data
      • insufficient variance
      • dataset layout
    • Data Engineering
      • PIL
      • Normalization and Standardization
      • label encoding
      • data splitting
      • openCV
    • Data Augmentation
      • under-fitting
      • perspective
      • flipping
      • rotation
      • code lab with Keras
    • Data Curation
      • population distribution
      • sampling distribution
      • code lab with Keras
    Day 2 (9am -4:30pm)
    Module 3: Training Models
    • Training Preparation
      • splitting
      • shuffling
      • stratification
      • code lab
    • Hyperparameter Tuning
      • Epochs and Steps
      • Batch size
      • learning rate
      • optimizer
      • feeding
      • code lab
    • Training
      • pre-training
      • weight initialization
      • Grid Search
      • Gradient Descent
    • Pre-Built Models & Transfer Learning
    Module 4: Deployment and Production
    • Intro to TensorFlow 2.0, tf.Keras, tools (Colab, TensorBoard)
    • Deploy models with TensorFlow serving
    • Model training and deployment in the browser with TensorFlow JS
    • On-device ML: train a model from scratch, convert to TFLite and deploy to mobile and IoT
    • Demo of TFLite models on microcontroller and Coral Edge TPU
    12-1:30pm: Lunch break
    Andrew Ferlitsch

    Andrew is a machine learning expert at Google. he educates software engineers in machine learning and AI. He is the creator of and oversees the development of the open source project Gap, which is a ML data engineering framework for computer vision.
    Margaret Maynard-Reid
    Margaret is a Google Developer Expert for machine learning who develops apps with intelligence. She leads GDG Seattle and co-organizes Seattle Data/Analytics/Machine Learning. She writes blogs and speaks at conferences on TensorFlow, deep learning and Android. She is passionate about community building and helping others get into AI/ML.
    • Start Date: ended
    • Venue: Galvanize, Seattle
    • Fee:
      $399 $399
    • Students enrolled:60
    • Status: course ended
    • Preview this course:
    Enroll This Course