“A first look at Pytorch 1.0” by @fastdotai alum @nikhilbalaji https://t.co/3ynMjlq8sY
— Jeremy Howard (@jeremyphoward) October 12, 2018
This is the best overview of the new @Pytorch 1.0 features and direction I've seen anywhere.
“A first look at Pytorch 1.0” by @fastdotai alum @nikhilbalaji https://t.co/3ynMjlq8sY
— Jeremy Howard (@jeremyphoward) October 12, 2018
This is the best overview of the new @Pytorch 1.0 features and direction I've seen anywhere.
Vel: PyTorch meets (OpenAI) baselines
— PyTorch (@PyTorch) October 11, 2018
A post on the Vel package which has a large pool of well-test pre-built baseline components for RL and Vision.
The debugging stories while building the package are valuable for researchers and engineers.https://t.co/udJSfiK9xg pic.twitter.com/9L3F6ElHtW
At the @PyTorch developer conference, I was part of a fascinating panel with @clattner_llvm, Yangqing Jia, and Noah Goodman, expertly moderated by @soumithchintala . Here it is!https://t.co/oPY1bsPZRO
— Jeremy Howard (@jeremyphoward) October 10, 2018
How to Use @PyTorch 1.0 and fastai 1.0 on Google Compute Engine (with pre-installed tutorials for both!) @googlecloud https://t.co/N8Lf84hudw
— Rachel Thomas (@math_rachel) October 7, 2018
It's not at all a hobby. Facebook has put serious engineering muscles behind PyTorch. It's used to train production systems, so it can't be just a hobby .
— Yann LeCun (@ylecun) October 6, 2018
But one should keep in mind that PyTorch is an open source project, not a Facebook-only project.
Thanks @Barzhini for talking to me about our newly released fastai v1 library for @PyTorch, on @ZDNet video. Includes a discussion of recent research from @mcleavey at @OpenAIhttps://t.co/in0rD38tkI
— Jeremy Howard (@jeremyphoward) October 4, 2018
PyTorch 1.0 is here!https://t.co/Kr0UQq16EF https://t.co/Kr0UQq16EF
— Yann LeCun (@ylecun) October 2, 2018
Google Cloud to support @PyTorch 1.0 on:
— hardmaru (@hardmaru) October 2, 2018
- Deep Learning VM Images
- Kubeflow
- TensorBoard integration
- Cloud TPUshttps://t.co/QAM8G1ANqX
Nvidia Vid2Vid for high-res video-to-video translation @ctnzr #PyTorch https://t.co/a3G55LjO2K
— Rachel Thomas (@math_rachel) October 2, 2018
fastai v1, built on @PyTorch v1 is now live (https://t.co/vkcy1YFySh). Awesome work by @jeremyphoward and @GuggerSylvain! This is now the fastest way to fine-tune a language model. https://t.co/PtBH3HlOkG
— Sebastian Ruder (@seb_ruder) October 2, 2018
Announcing #PyTorch for Google TPUs! Prototype training of ResNet 50 right now. I've said it before-- I'm looking this collaborative spirit! pic.twitter.com/cXU7XT0wrJ
— Rachel Thomas (@math_rachel) October 2, 2018
Google’s @aronchick at @PyTorch conference: “We’re happy to announce Google is collaborating with PyTorch to integrate with TensorBoard and Cloud TPUs” pic.twitter.com/Flk4MjdHE4
— Peter Skomoroch (@peteskomoroch) October 2, 2018