We recently released Autograd for Torch, which greatly simplified our workflow when experimenting with complex deep learning architectures. The Twitter Cortex team is continuously investing in better tooling for manipulating our large datasets, and distributing training processes across machines in our cluster.
Today we’re open-sourcing four components of our training pipeline, so the community using Torch and/or Autograd can simplify their workflows when it comes to parallelizing training, and manipulating large, distributed datasets.
When the path-breaking American astronaut Neil Armstrong passed away at age 82 on August 25th, Twitter lit up as people around the world shared the news. Within about a day, their conversations and condolences added up to 1.6 million Tweets.
We often say “You never know where Twitter will take you,” but here’s one that (almost) defies imagination. Last night, we saw that NASA’s Jet Propulsion Laboratory (@NASAJPL) was live-tweeting the journey to Mars. As Curiosity, the car-sized rover, made its descent onto the Red Planet, the #JPL team shared Twitter updates directly from the command center. And that stream reveals a truly remarkable and historic sequence of events:
Yesterday, around the world, people paused their daily routines to gaze upward for a glimpse of the annular solar eclipse. They pointed cameras and cameraphones towards the sun or at the deepening dancing shadows in hopes of capturing images of this incredible sight. Collectively, these photos transcend language to tell the story of people marveling at one of our solar system’s most magical moments.
Last week we kicked off a new blog series showcasing the amazing array of moments that occur #OnlyOnTwitter. We’re back again with highlights from this past week, which take Twitter to the deepest part of the ocean and a racecar driver’s seat in the sky.