#PyTorch re-implementation of DeepMind's Perceiver IO: A General Architecture for Structured Inputs & Outputs https://t.co/c16ftYKgzJ pic.twitter.com/srwT1TiOaU
— Alexandr Kalinin (@alxndrkalinin) August 30, 2021
#PyTorch re-implementation of DeepMind's Perceiver IO: A General Architecture for Structured Inputs & Outputs https://t.co/c16ftYKgzJ pic.twitter.com/srwT1TiOaU
— Alexandr Kalinin (@alxndrkalinin) August 30, 2021
What’s more, if your training is in @PyTorch, you can rather easily add this behaviour with minimal changes to your codebase, using @higherpytorch.https://t.co/U5dFLBXTHZ
— Edward Grefenstette 🇪🇺 (@egrefen) August 28, 2021
It's time to stop making t-SNE & UMAP plots. In a new preprint w/ Tara Chari we show that while they display some correlation with the underlying high-dimension data, they don't preserve local or global structure & are misleading. They're also arbitrary.🧵https://t.co/XkAOTKlOcs pic.twitter.com/dmFzD5RR6R
— Lior Pachter (@lpachter) August 27, 2021
One TTS Alignment To Rule Them All
— AK (@ak92501) August 25, 2021
pdf: https://t.co/yQu2GWB6uw
abs: https://t.co/cAAxzOuGGg
present an alignment framework that is broadly applicable to various TTS architectures, both autoregressive and parallel pic.twitter.com/e4tpCSNwWC
How Can Increased Randomness in Stochastic Gradient Descent Improve Generalization?
— AK (@ak92501) August 24, 2021
pdf: https://t.co/Jsj1hpi3vB
abs: https://t.co/nEGOGZ2Z8v pic.twitter.com/weTbqMbQbF
Our RemBERT model (ICLR 2021) is finally open-source and available in 🤗 Transformers.
— Sebastian Ruder (@seb_ruder) August 23, 2021
RemBERT is a large multilingual Transformer that outperforms XLM-R (and mT5 with similar # of params) in zero-shot transfer.
Docs: https://t.co/AKwV0UF6cT
Paper: https://t.co/TXF7qlJtUY pic.twitter.com/ytIiMOqVks
Do Vision Transformers See Like Convolutional Neural
— AK (@ak92501) August 20, 2021
Networks?
pdf: https://t.co/5Yz5F2PZwO
abs: https://t.co/bpHO2rOYDv
find striking differences between the two architectures, such as ViT having more uniform representations across all layers pic.twitter.com/0KT0KE16f9
"Pitfalls in Machine Learning Research: Reexamining the Development Cycle" -- Nice write-up by Stella Biderman & Walter Scheirer on improving your methodological ML practices, incl (1) algorithm design/choice, (2) data collection, and (3) model evaluation: https://t.co/YMcCFhpWCv
— Sebastian Raschka (@rasbt) August 18, 2021
Real-ESRGAN: Training Real-World Blind Super-Resolution with Pure Synthetic Data now on @huggingface Spaces using @Gradio
— AK (@ak92501) August 17, 2021
demo: https://t.co/3KXzlj5M7Z
paper: https://t.co/1j8gR7uDIC
github: https://t.co/IE0O5SHDDF pic.twitter.com/4LH9oIWO15
SOTR: Segmenting Objects with Transformers
— AK (@ak92501) August 17, 2021
pdf: https://t.co/eplIKD4mgZ
abs: https://t.co/ARAaQ7VJAe
github: https://t.co/XlVZrJh25P
performs well on the MS COCO dataset and surpasses sota instance segmentation approaches pic.twitter.com/06tH3XPtKQ
There are now many ways to learn about DALL·E mini, the text to image generator 🎉
— Boris Dayma 🥑 (@borisdayma) August 13, 2021
📹 for viewers, see the presentation: https://t.co/eGV22uhBn0
📕 for readers, see the report: https://t.co/liX9qN79hj
🥑 play with the demo: https://t.co/OiBcNrqoBv
This is a really great NLP Transformer survey, indeed! Also, I like that they included a section focusing on the three main ways to utilize a pre-trained transformer (assuming most of us don't have the infrastructure to train them from scratch): https://t.co/lGULN6vCwV https://t.co/HGRJoGmISM pic.twitter.com/UEHL0MSPwo
— Sebastian Raschka (@rasbt) August 13, 2021