We release LLM.int8(), the first 8-bit inference method that saves 2x memory and does not degrade performance for 175B models by exploiting emergent properties. Read More:
— Tim Dettmers (@Tim_Dettmers) August 17, 2022
Paper: https://t.co/eNpinXS0Z5
Software: https://t.co/hBuVyQhLqS
Emergence: https://t.co/oPGRhACNEe pic.twitter.com/vNWxrDHlOh
“Giant squid assembling IKEA furniture”
— hardmaru (@hardmaru) August 17, 2022
These weird images are generated by combining *both* #DalleMini🖍️ and #StableDiffusion🦑 models‼️
🧵 I’ll explain how to do it in this thread 👇 pic.twitter.com/E2EGBrLPZh
Here's a Pandas trick to add order to categorical values. 🤯🐼 pic.twitter.com/8JD4AkfEzS
— Matt Harrison (@__mharrison__) August 13, 2022
The Man behind #StableDiffusion — An Interview with Emad Mostaque (@EMostaque), founder of Stability AI.
— hardmaru (@hardmaru) August 13, 2022
Really inspiring interview.https://t.co/Ham198RSm8
A Python-level JIT compiler designed to make unmodified #PyTorch programs faster. https://t.co/bZrmqtxCbO #deeplearning #machinelearning #ml #ai #neuralnetworks #datascience
— PyTorch Best Practices (@PyTorchPractice) August 13, 2022
Can one deliver a RecSys masterclass with a focus on
— Radek Osmulski 🇺🇦 (@radekosmulski) August 12, 2022
• online vs batch predictions
• monitoring: data distribution shifts
• model deployment
in 25 minutes? 🤔
Apparently, @chipro can! 🙂 A must-watch
Link to video: pic.twitter.com/gWWUI7WGbc
I really enjoyed this GPU programming course:https://t.co/eYmqzrk1eq
— Jeremy Howard (@jeremyphoward) August 11, 2022
enjoying this awesome tutorial on conformal prediction by @ml_angelopoulos & @stats_stephen: https://t.co/TBuVvn27yC. what a nice, straightforward framework to think of UQ in practical terms.
— Kyunghyun Cho (@kchonyc) August 10, 2022
there’s an associated tutorial paper as well: https://t.co/76PNfSIYlJ.
Just saw this comprehensive thesis on deep learning for OOD data drop on arXiv: "Towards Robust Deep Learning using Entropic Losses" https://t.co/MKXq6hbwWZ.
— Sebastian Raschka (@rasbt) August 10, 2022
Lots to read here, but in a nutshell, it looks like the proposed IsoMax+ performs well and is relatively easy to use 👌 pic.twitter.com/aF3Kpyd6QN
Atlas: a not-so-large language model (11B parameters) that beats the big guys at question answering and fact checking.
— Yann LeCun (@ylecun) August 8, 2022
The main difference is that it can retrieve facts from a corpus.
Paper: https://t.co/WlV3rbuY30 https://t.co/7bsNXfqV6l
Another common myth: the number of training data points has to be larger than the number of model parameters.
— Sebastian Raschka (@rasbt) August 8, 2022
Btw, the full code is here: https://t.co/KYYnlB4fn3
AI really is going to change the world. Things are going to get 100-1000X cheaper and more efficient. This is mostly great. However, historically, when you make stuff 100X-1000X cheaper, you upend the geopolitical order. This time probably won't be different.
— Jack Clark (@jackclarkSF) August 6, 2022
