David Solís (@davsolmar) profile picture

Father, partner, computer scientist, data scientist, entrepreneur, profesor, dancer, rolling blader, climber, ...

Sevilla, España
Joined May 2013
Photo by David Solís (@davsolmar)Photo by David Solís (@davsolmar)Photo by David Solís (@davsolmar)Photo by David Solís (@davsolmar)Photo by David Solís (@davsolmar)Photo by David Solís (@davsolmar)Photo by David Solís (@davsolmar)Photo by David Solís (@davsolmar)Photo by David Solís (@davsolmar)
Pues eso, que he estado probando la IA esta que traduce directamente a cualquier idioma y creo que ya vivimos en 2080
90
721
143
2,241
310,925
David Solís retweeted
Cuando llegaba la tormenta a Isla Cristina. Fotografías de Antonio G. Alonso (facebook.com/antonio.alonso.…)
21
131
19
566
66,398
David Solís retweeted
Cool work, have a look! Interesting to see they tie the "probability" of discrete representation to, well, the probability of the dataset : Variational Inference itself. arxiv.org/abs/2304.04820
22
3
103
12,484
David Solís retweeted
WTF: Mind reading is here. Researchers invented a new #AI method to convert brain signals into video. See the results for yourself Published in Nature yesterday: nature.com/articles/s41586-0… What are the implications? Is this the biggest paper of 2023? #CEBRA
118
783
220
3,054
679,605
Result: quality similar to Copilot at launch! - Outperform every open code LLM that supports multiple programming languages - Matches/outperforms Copilot at launch - 40% pass@1 on HumanEval
2
5
28
6,922
Tracking through Containers and Occluders in the Wild Proposes a model, called TCOW, that can segment objects in videos with a notion of object permanence. proj: tcow.cs.columbia.edu/ abs: arxiv.org/abs/2305.03052
1
15
1
37
6,562
David Solís retweeted
How to summarize a book without sending 100% of your tokens to a LLM I tested out a fun method on "Into Thin Air" (~140K tokens) The results were surprising good Definitely enough to refine and keep working Here's a speed run though process:
37
88
14
738
227,610
ZipIt! Merging Models from Different Tasks without Training Presents ZipIt!, a general method for merging two arbitrary models of the same architecture that incorporates two simple strategies. arxiv.org/abs/2305.03053
3
81
5
396
43,573
Today we are releasing a CLIP ViT-L/14 model with 79.2% zero-shot accuracy on ImageNet. Our model outperforms OpenAI's CLIP by a large margin, and outperforms even bigger models (ViT-g/14) trained on LAION-2B Check it out at huggingface.co/laion/CLIP-Vi…!
16
126
14
704
191,062
BREAKING: The future of object detection is here! Meet YOLO-NAS, open-sourced YOLO-based architecture redefining state-of-the-art in object detection.
60
587
97
3,044
726,704
What happens when you divide by zero on a mechanical 1950s calculator
854
6,570
730
60,258
7,361,939
David Solís retweeted
Real-Time Radiance Fields for Single-Image Portrait View Synthesis abs: arxiv.org/abs/2305.02310 project page: research.nvidia.com/labs/nxp…
4
63
7
336
225,124
David Solís retweeted
Leaked Google document: “We Have No Moat, And Neither Does OpenAI” The most interesting thing I've read recently about LLMs - a purportedly leaked document from a researcher at Google talking about the huge strategic impact open source models are having simonwillison.net/2023/May/4…
113
1,114
255
4,787
1,944,900
David Solís retweeted
Correlación entre consumo de chocolate por países y ganadores del Nobel. Maravilloso.
My favourite science policy slide: strong correlation between Nobel prizes and chocolate consumption.
True story.
124
335
112
2,420
Este trolleo a la administración es de quitarse el sombrero, a ver si así se dieran cuenta de los que los usuarios pasamos continuamente
6
110
3
336
David Solís retweeted
stylegan3-projector Mario github: github.com/ouhenio/stylegan3…
60
336
106
2,040
David Solís retweeted
The first 2 weeks of the 2021 edition of the NYU Deep Learning course are now on line, including lectures, practicum, homework, etc.
📣 NYU Deep Learning SP21 📣 Theme 1 / 3: Introduction Video lectures, notebooks, slides, and HOMEWORK have been released! 🥳🥳🥳 Website: atcold.github.io/NYU-DLSP21/ Lecture 1: youtu.be/nTlCqaL7fCY Lecture 2: youtu.be/IYQN3i7dJIQ
4
54
1
334
David Solís retweeted
Climate Change: A Timeline :-)(
30
1,020
68
3,492
Multimodal Few-Shot Learning with Frozen Language Models Present a simple, yet effective, approach for transferring the few-shot learning ability of LM to a multimodal setting (vision and language) by combining a frozen LM with a trained visual encoder. arxiv.org/abs/2106.13884
1
11
2
99