Excellent work on a flexible framework for handling nonstandard interpretations in probabilistic programming. For example, different inference strategies to estimate integrals in a composable manner. Was a nice read! https://t.co/OT8b5iqeph
— Dustin Tran (@dustinvtran) October 25, 2019
For those interested in following along with all of my notebooks for the @rsna @kaggle competition, I've made a little summary of the whole process.
— Jeremy Howard (@jeremyphoward) October 25, 2019
I hope you find this helpful! :)https://t.co/4gKxdAE5IT pic.twitter.com/zSqciy5NgS
TensorSketch! https://t.co/GnryzBPctc
— hardmaru 😷 (@hardmaru) October 25, 2019
EagerPy is a thin wrapper around PyTorch and TensorFlow Eager that unifies their interface and thus allows writing code that works with both. https://t.co/ogUjDgR7T3 #deeplearning #machinelearning #ml #ai #neuralnetworks #datascience #tensorflow
— TensorFlow Best Practices (@TFBestPractices) October 24, 2019
Yay! Figured out how to do `groupby` in @pytorch. It's not pretty, but it is fast pic.twitter.com/mFkL0mzOLB
— Jeremy Howard (@jeremyphoward) October 18, 2019
How to Run @PyTorch with TPUshttps://t.co/j31nrzROAX
— hardmaru 😷 (@hardmaru) October 18, 2019
PyTorch 1.3 now supports @GCPcloud TPUs (public alpha). pic.twitter.com/0DcWVItlad
Captum – model interpret-ability & understanding library for PyTorch
— ML Review (@ml_review) October 17, 2019
Includes a web interface for easy visualization and access to a number of interpretability algorithms: Integrated Gradients, Saliency Maps, Smoothgrad, Vargrad and othershttps://t.co/QHPlf29n1v pic.twitter.com/bSmDpuZV2u
We're extremely excited to release our BoTorch paper!
— Andrew Gordon Wilson (@andrewgwils) October 16, 2019
Scalable, flexible, and modular Bayesian optimization integrated with GPyTorch and @PyTorch. It's been a pleasure working with @eytan, Max, and team.
paper: https://t.co/NUe6ve4i8p
code: https://t.co/RyEp1NneS4 pic.twitter.com/jc18vkEQ12
Yes! Comments shouldn’t have to be used as a crutch for lack of clarity in code. This is a very positive step forward. So much time can be wasted chasing bugs that stem purely from confusion. https://t.co/gqwEZ7LYGD
— Jason Antic (@citnaj) October 11, 2019
New semantic (more human-readable) naming schemes for linear algebra in PyTorch 1.3#PTDC2019 @PyTorch pic.twitter.com/oPTeAdPcAb
— Rachel Thomas (@math_rachel) October 10, 2019
PyTorch 1.3 includes support for model deployment to mobile devices, quantization, & front-end improvements, like the ability to name tensors. New tools & libraries are also launching for improved model interpretability & multimodal development. Read more: https://t.co/HaaIt9cU3U pic.twitter.com/UKHdU7sHIH
— Facebook AI (@facebookai) October 10, 2019
Announcing TorchBeast, an IMPALA-inspired @pytorch platform for distributed RL research. Used in a growing number of projects here at @facebookai. Project lead by Heinrich Küttler, with major effort by @nntsn et al.
— Edward Grefenstette (@egrefen) October 9, 2019
Paper: https://t.co/HMYTtesSqC
Code: https://t.co/vadpUsyvaZ
