Explore an end-to-end deep learning pipeline for self-driving cars in this 44-minute conference talk. Dive into the world of autonomous vehicles, starting with a toy car implementation and progressing through data collection, processing, and model training using TensorFlow. Learn about the pipeline components, including DonkeyCar, data serialization, and neural network architecture. Discover common pitfalls, evaluation metrics, and deployment strategies. Gain insights into scaling up to Level 5 autonomous vehicles and the importance of simulation in development. Understand the challenges of building self-driving cars and how to track results effectively in software projects.
End-to-End Deep Learning Pipeline for Self-Driving Cars
Overview
Syllabus
Introduction
Selfdriving cars
Selfdriving toy car
Deep learning
Outline
Pipeline
DonkeyCar
Data collection
Getting data
Data collection pipeline
Data points
Data processing
Data splitting
Motivation
Tensorflow records
Serialization
sterilized mode
pictures
deep layers
input
convolution
max pooling
dense layer
dropout layer
training
training script
evaluation
metrics
measurement
typical mistakes
deployment
summary
driving errors
all software projects
tracking results
running the model
data gathering
building a selfdriving car
level 5 selfdriving cars
validation and test dates
why Tensorflow
how many people
simulation
Taught by
WeAreDevelopers
Related Courses
-
Machine Learning & Self-Driving Cars: Bootcamp with Python
-
Visual Perception for Self-Driving Cars
-
Data Science: Deep Learning Project for Self Driving Cars
-
Computer Vision and Perception for Self-Driving Cars (Deep Learning Course)
-
Inside NVIDIA’s AI Infrastructure for Self-Driving Cars
-
Machine Learning for Autonomous Vehicles
