Vision Permutator: A Permutable MLP-Like Architecture for Visual Recognition
— AK (@ak92501) June 24, 2021
pdf: https://t.co/2Kb6zoFwR1
github: https://t.co/Ahe17BEDwf
achieves 81.5% top-1 accuracy on ImageNet without extra large-scale training data (e.g., ImageNet-22k) using only 25M learnable parameters pic.twitter.com/KO1Kx7OnSi
Despite its popularity in the covariate shift setting, Bayesian model averaging can surprisingly hurt OOD generalization! https://t.co/AqbH4f29FG 1/5 https://t.co/OawZs6AV7n pic.twitter.com/SiPHp8mQrm
— Andrew Gordon Wilson (@andrewgwils) June 23, 2021
Revisiting Deep Learning Models for Tabular Data
— AK (@ak92501) June 23, 2021
pdf: https://t.co/G4J8TRyRBt
abs: https://t.co/tMUhp2IZwW
github: https://t.co/qeue47mnWD
proposed a attention-based architecture that outperforms ResNet on many tasks pic.twitter.com/aBCZ3dD4VV
There are three types of researchers: 1. those that only look at the paper's methods and ideas; 2. those that only look at experiments; and 3. those that no longer read papers.
— Dustin Tran (@dustinvtran) June 23, 2021
"Paraphrastic representations at scale" is a strong, blazing fast package for sentence embeddings by @johnwieting2
— Graham Neubig (@gneubig) June 22, 2021
Paper: https://t.co/LlHbH788b2
Code: https://t.co/91XR75PDG9
Beats Sentence-BERT, LASER, USE on STS tasks, works multilingually, and is up to 6,000 times faster 😯 pic.twitter.com/julWhnjyYf
Regularization is all you Need: Simple Neural Nets can Excel on Tabular Data
— AK (@ak92501) June 22, 2021
pdf: https://t.co/rcRbf6H3tf
abs: https://t.co/voMqog8zct
well-regularized plain MLPs significantly outperform sota specialized neural network architectures pic.twitter.com/L3jcM0hbgc
What if we could mutualize everyone's compute? 🤯🤯🤯
— Hugging Face (@huggingface) June 21, 2021
Super proud to share the result of our first experiment of decentralized collaborative training with 40 volunteers!
The result is an Albert-like model for Bengali competitive with the current SOTA!
📰 https://t.co/kdiuPmpaPv pic.twitter.com/ymEmU6UtYU
World-GAN: a Generative Model for Minecraft Worlds
— AK (@ak92501) June 21, 2021
pdf: https://t.co/rSx646L1hp
abs: https://t.co/RJcHEOTUQT pic.twitter.com/n0sw8fEAWF
Efficient Self-supervised Vision Transformers for Representation Learning
— AK (@ak92501) June 21, 2021
pdf: https://t.co/0YImgGI4Fg
abs: https://t.co/gIxv7H8Qsn
EsViT achieves 81.3% top-1 on the ImageNet linear
probe evaluation, outperforming prior arts with around an order magnitude of higher throughput pic.twitter.com/28P5rH7AW0
How to train your ViT? Data, Augmentation, and Regularization in Vision Transformers
— AK (@ak92501) June 21, 2021
pdf: https://t.co/6isx98PS1M
train ViT models of various sizes on the
ImageNet-21k dataset, either match or outperform their counterparts trained on larger, but not publicly available JFT-300M pic.twitter.com/Zoso6tUPPA
It’s important to identify the common elements of all the failed #datascience projects you’ve ever worked on so you can work on removing them.
— Nihilist Data Scientist (@nihilist_ds) June 20, 2021
Of course, the biggest one is that they all involved you. #rstats #pydata
It’s never been easier to deploy a state-of-the-art ML model to a phone. In this blog, we’ll provide a quick overview of PyTorch Mobile-powered demo apps running various state-of-the-art PyTorch 1.9 ML models spanning images, video, audio and text.https://t.co/RxNNnktYvr
— PyTorch (@PyTorch) June 18, 2021
