Mobility – Page 6910 – I wish I had more to offer

Tesla’s Andrej Karpathy Talks PyTorch, Autopilot (Video)

Invest
Electric Cars
Electric Car Benefits
Electric Car Sales
Solar Energy Rocks
RSS
Advertise
Privacy Policy

Autonomous Vehicles

Published on December 5th, 2019 |

by Guest Contributor

Tesla’s Andrej Karpathy Talks PyTorch, Autopilot (Video)

Twitter
LinkedIn
Facebook

December 5th, 2019 by Guest Contributor

Originally published on EVANNEX.
By Charles Morris

Andrej Karpathy, Tesla’s Director of Artificial Intelligence and Autopilot Vision, is one of the chief architects of Tesla’s self-driving vision. In July, he hosted a workshop on Neural Network Multi-Task Learning, where he offered some detailed insights on Tesla’s use of AI in developing its Autopilot features.

Now Karpathy is featured in a new video in which he describes how Tesla is using PyTorch, an open-source machine learning library, to develop full self-driving capabilities for its vehicles, including Navigate on Autopilot and Smart Summon.

A look at Tesla’s Autopilot (Image: Tesla)

Karpathy explains that, unlike other companies working on self-driving, Tesla doesn’t use lidar or high-definition maps, so the Autopilot system relies on AI to parse information from the eight cameras mounted around the vehicle. Tesla is a fairly vertically integrated company, so it has control of the “full stack” when it comes to AI. The machine learning process is built around “hydranets,” so called because each has a shared backbone and multiple heads (like the Hydra of Greek mythology). Karpathy demonstrates how Tesla’s Smart Summon feature uses hydranets to figure out how to negotiate a parking lot.

Andrej Karpathy discusses the development of Tesla’s Autopilot and Smart Summon features (YouTube: PyTorch)

Karpathy’s talk gets very technical very quickly — only those with a background in machine learning are likely to be able to follow the full story here. However, even we laypeople can appreciate the incredible complexity of teaching a computer to drive a car. According to Karpathy, compiling a full build of Autopilot 2.0 involves some 48 different networks, 1,000 distinct predictions, and 70,000 GPU hours. And, of course, this is no one-time project — the software is continuously being improved, so it must be frequently re-compiled and updated.

This continuous improvement is driven by the massive amounts of data pouring in from the fleet of Teslas on the world’s roads — an asset no other company working on autonomous driving enjoys. Karpathy tells us that the Navigate on Autopilot feature has now accumulated over a billion miles of real-world usage in over 50 countries, including 200,000 automated lane changes. The Smart Summon feature has been used in over 500,000 sessions in the short time since it was introduced. Keep this figure in mind the next time some pundit declares Smart Summon a failure because of a handful of YouTube videos of comical parking lot mishaps.
Follow CleanTechnica on Google News.
It will make you happy & help you live in peace for the rest of your life.

About the Author

Guest Contributor is many, many people. We publish a number of guest posts from experts in a large variety of fields. This is our contributor account for those special people. 😀

Advertise with CleanTechnica to get your company in front of millions of monthly readers.

@GM: General Motors and LG Chem Team Up to Advance Toward an All-Electric Future, Add Jobs in Ohio DETROIT , Dec. 5, 2019 /PRNewswire/ — Will create mo…

DETROIT, Dec. 5, 2019 /PRNewswire/ — Will create more than 1,100 new jobs in a new, Ohio-based facility Joint venture to drive cost per kilowatt hours to industry-leading levels New battery plant to have an annual capacity of more than 30 gigawatt hours with flexibility for expansion Companies enter joint agreement to develop advanced tech,… Continue reading @GM: General Motors and LG Chem Team Up to Advance Toward an All-Electric Future, Add Jobs in Ohio DETROIT , Dec. 5, 2019 /PRNewswire/ — Will create mo…

@Hyundai: Hyundai Motor Unveils ‘Strategy 2025’ Roadmap

Hyundai to transition into Smart Mobility Solution Provider by 2025 with two pillars Smart Mobility Device will expand beyond automobiles to include Personal Air Vehicle (PAV), robotics, last-mile mobility Smart Mobility Service will offer personalized services and contents on integrated platform Hyundai announces a set of strategic goals achieve balanced and steady growth boost profitability… Continue reading @Hyundai: Hyundai Motor Unveils ‘Strategy 2025’ Roadmap

Waymo robotaxi app arrives on the App Store

Hailing one of Waymo’s self-driving minivans is about to get a little easier. One year after launching its Waymo One self-driving car service in the Phoenix area, the company is launching an app on iOS, the latest signal that the company is inching toward large-scale commercial service. Now, Phoenix residents can download the Waymo app… Continue reading Waymo robotaxi app arrives on the App Store

Einride to launch commercial pilot of driverless electric pods with Coca-Cola European Partners

Autonomous robotic road-riding cargo pod startup Einride has signed a new partner for a commercial pilot on Sweden’s roads, which should be a great test of the company’s electric driverless transportation pods. Einride will be providing service for Coca-Cola European Partners, which is the official authorized bottler, distributor, sales and marketing company for Coca-Cola branded… Continue reading Einride to launch commercial pilot of driverless electric pods with Coca-Cola European Partners

A look at Latin America’s emerging fintech trends

Thiago Paiva Contributor Thiago is a fintech entrepreneur, investor, and columnist. He is currently a product leader at Oyster, a neobank for SMEs in Latin America. More posts by this contributor The battle to become the Mexican Nubank just started The emergence of super apps in Latin America Although the 2008 global financial crisis sparked… Continue reading A look at Latin America’s emerging fintech trends

@VW Group: In brief: Key components for a new era – the battery system

Up to 550 kilometres with just one battery charge: Volkswagen’s fully electric ID.3 has impressive driving dynamics and great ranges. It is the first model of a product line of vehicles based on the modular electric drive matrix (MEB). Its fast charging capability allows the ID.3 to be recharged with 100 kW charging power within… Continue reading @VW Group: In brief: Key components for a new era – the battery system

@VW Group: Where Innovation happens: Experts discuss ideas for future mobility at the MQ! Innovation Summit in Beijing

Under the motto #neverstopquestioning, Audi has gathered pioneers, visionaries and technical experts from all over the world in Beijing for the third MQ! Innovation Summit. The core question remains: Is there a mobility quotient (MQ), similar to an IQ? In what way could the MQ be measured? How will sustainable mobility be defined in the… Continue reading @VW Group: Where Innovation happens: Experts discuss ideas for future mobility at the MQ! Innovation Summit in Beijing

Hyundai Strategy 2025: 670,000 BEVs/FCEVs Annually By 2025

Hyundai bets on mostly battery-electric cars in its new strategic plan. Hyundai recently announced a bold business plan – “Strategy 2025” – which outlines the goals for 2025 and beyond. Electrification is one of the three strategic directions. The South Korean brand aims to become one of the top three EV manufacturers by 2025. Sales… Continue reading Hyundai Strategy 2025: 670,000 BEVs/FCEVs Annually By 2025

Lyft: Leveraging Early Sensor Fusion for Safer Autonomous Vehicles

By: Interns Eric Vincent¹ and Kanaad Parvate², with support from Andrew Zhong, Lars Schnyder, Kiril Vidimce, Lei Zhang, and Ashesh Jain from our Perception Team Accurate perception to enable safe autonomy In order to safely navigate the world, an autonomous vehicle (AV) needs to accurately perceive the objects that surround it. This includes dynamic objects such… Continue reading Lyft: Leveraging Early Sensor Fusion for Safer Autonomous Vehicles