Blaz Skrlj

597 posts

Blaz Skrlj banner
Blaz Skrlj

Blaz Skrlj

@SkBlaz

🦀

Europe Katılım Ağustos 2017
555 Takip Edilen83 Takipçiler
Blaz Skrlj retweetledi
Cameron R. Wolfe, Ph.D.
Cameron R. Wolfe, Ph.D.@cwolferesearch·
The ChatGPT API was released yesterday and it costs 90% less than expected. Here’s five methods (and resources to learn about them) that are **probably** being used to enable this price reduction… 🧵[1/6]
Cameron R. Wolfe, Ph.D. tweet media
English
25
245
1.3K
263.2K
Blaz Skrlj retweetledi
Ching-Lung Hsu
Ching-Lung Hsu@hiallen72·
Hippocampal spatial representations exhibit a hyperbolic geometry that expands with experience | Nature Neuroscience nature.com/articles/s4159…
English
6
114
489
114.8K
Blaz Skrlj retweetledi
Mark Tenenholtz
Mark Tenenholtz@marktenenholtz·
We've made so much progress in the last 7-8 years on complex domains like text and images and basically none on simple sequences of numbers (i.e. time-series forecasting). In fact, I think we've regressed since good information is so hard to find.
English
21
18
230
80.4K
Blaz Skrlj retweetledi
vicki
vicki@vboykis·
GitHub just wrote an article about how they had to write their own search engine (in Rust, for performance reasons) and a new probabilistic data structure to reduce indexing time because ES and Lucene were blockers for them but sure big data is over 😅
vicki tweet media
English
29
95
1.1K
198.9K
Blaz Skrlj retweetledi
DJ
DJ@DuaneJRich·
One day I'll get to studying Normalizing Flows. When I do, I'll start here github.com/janosh/awesome…
DJ tweet media
English
6
127
691
68.7K
Blaz Skrlj retweetledi
George E. Dahl
George E. Dahl@GeorgeEDahl·
We've just released the first version of our Deep Learning Tuning Playbook! This is our attempt to distill our process for actually getting good results with deep learning. We emphasize hyperparameter tuning since it has been a large pain point. github.com/google-researc…
English
44
794
3.6K
671.2K
Blaz Skrlj retweetledi
Fermat's Library
Fermat's Library@fermatslibrary·
23456789 is the largest prime number with consecutive increasing digits
English
33
171
1.3K
0
Blaz Skrlj retweetledi
Guodong Zhang
Guodong Zhang@Guodzh·
How to train very deep NNs without shortcuts, but still achieve competitive results on ImageNet? Our ICLR paper gives a simple solution derived from kernel approx theory. We hope this could enable further research into deep models. openreview.net/pdf?id=U0k7XNT…
Guodong Zhang tweet media
English
6
48
347
0
Blaz Skrlj retweetledi
Martina Maritan
Martina Maritan@MartinaMaritan·
animation work in progress🤓 - went from sphere to actual 3D Whole-Cell model #unity
English
6
50
323
0
Blaz Skrlj retweetledi
Russell Kaplan
Russell Kaplan@russelljkaplan·
Today I saw the impact that AlphaFold is having on speeding up drug discovery firsthand:
English
16
526
3.3K
0
Blaz Skrlj retweetledi
Colin McCarthy
Colin McCarthy@US_Stormwatch·
Tonga's Hunga Tonga volcano just had one of the most violent volcano eruptions ever captured on satellite.
GIF
English
1.2K
38.7K
108.8K
0
Blaz Skrlj retweetledi
AK
AK@_akhaliq·
GradMax: Growing Neural Networks using Gradient Information abs: arxiv.org/abs/2201.05125 a new approach for growing neural networks that aims to maximize the gradient norm when growing
AK tweet media
English
1
38
246
0