"Getting angry at people for making mistakes doesn't teach them not to make mistakes.
— Bojan Tunguz (@tunguz) September 11, 2021
It teaches them to hide their mistakes."
"Getting angry at people for making mistakes doesn't teach them not to make mistakes.
— Bojan Tunguz (@tunguz) September 11, 2021
It teaches them to hide their mistakes."
ConvMLP: Hierarchical Convolutional MLPs for Vision
— AK (@ak92501) September 10, 2021
pdf: https://t.co/f6c1XmyLSX
abs: https://t.co/vkgXvlCmcD
github: https://t.co/FyjNR8W3Oq pic.twitter.com/RXXJgzoula
Efficient Nearest Neighbor Language Models
— AK (@ak92501) September 10, 2021
pdf: https://t.co/uEtibkYA1L
abs: https://t.co/K0hBVhtvlk pic.twitter.com/ZKd0vnBDCs
Releases Python's GIL to allow use of multiple CPU cores—no need to use multiprocessing. And compatibility with JAX and @TensorFlow means it can be used with https://t.co/Fv77gxVyzu pipelines!
— 👩💻 Paige Bailey #BlackLivesMatter (@DynamicWebPaige) September 9, 2021
If you do ML/DL research in audio, make sure to check out this @Spotify library. 🎶🎹 https://t.co/WJfrT5OaIE pic.twitter.com/2ey4jxva7E
China is producing more steel than the rest of the world combined (and ten times more than the next largest country, India). Source: https://t.co/CflHYSLMzF pic.twitter.com/HsupvxAnGS
— Simon Kuestenmacher (@simongerman600) September 9, 2021
Scaled ReLU Matters for Training Vision Transformers
— AK (@ak92501) September 9, 2021
pdf: https://t.co/OcaIFc7Vfl
abs: https://t.co/1k2vrXXsoT pic.twitter.com/glmnhkahWj
Datasets: A Community Library for Natural Language Processing
— AK (@ak92501) September 8, 2021
abs: https://t.co/xEpY9oQ2a5
github: https://t.co/HvY6Nlf41c
650+ unique datasets, 250+ contributors, and has helped support a variety of novel crossdataset research projects and shared tasks pic.twitter.com/AdlB21Hu2c
I was never a good systems engineer, so I always avoided the topic of compiling and optimizing ML models.
— Chip Huyen (@chipro) September 8, 2021
However, as I work with ML on devices, the topic keeps coming up. So I spent the last 3 months learning about ML compilers.
Here’s what I learned. https://t.co/VNXbUcfJ2h
We're slowly learning more about Google's not-exactly-public efforts in the huge LM space. The highlight here for me was the subfigure on the right: More evidence that we can see discontinuous, qualitatively-important improvements in behavior as we scale.https://t.co/Q68HaF2Vag pic.twitter.com/yGHmwCLvPt
— Prof. Sam Bowman (@sleepinyourhat) September 7, 2021
PermuteFormer: Efficient Relative Position Encoding for Long Sequences
— AK (@ak92501) September 7, 2021
abs: https://t.co/S0bSxCDoc2
experiments show that PermuteFormer uniformly improves the performance of Performer with almost no computational overhead and outperforms vanilla Transformer on most of the tasks pic.twitter.com/PcmcRPrOtC
Deep Saliency Prior for Reducing Visual Distraction
— AK (@ak92501) September 7, 2021
pdf: https://t.co/ukJEBvPzH1
abs: https://t.co/voqLbUbPZs
project page: https://t.co/h2nvBRGgcD pic.twitter.com/7hNCqA0Ocb
Revisiting 3D ResNets for Video Recognition
— AK (@ak92501) September 7, 2021
pdf: https://t.co/UwhpllMj6z
abs: https://t.co/LofJjbdq2P
When pre-trained on a large Web Video Text dataset, our best model achieves 83.5% and 84.3% on Kinetics-400 and Kinetics-600 pic.twitter.com/F7Wmj2yW6X