Video yükleniyor...

Video Yüklenemedi

Ana Sayfaya Dön

What if you could induce models to be more concise via pretraining data curation?

leavittron's profile picture

Matthew Leavitt

3,387 subscribers

7,135,743 görüntüleme • 18 gün önce •via X (Twitter)

EğitimBilim & Teknoloji
LIVE

Anya Rossi• Live Now

Private livecam show

0 Yorum

Yorum bulunmuyor

Orijinal gönderinin yorumları burada görünecek

Benzer Videolar

Data curation is crucial for post-training recipes. But how do we curate? Curation is usually manual & tedious. And, it's hard to tell if a strategy in the data will be reliable! We introduce an automatic way to curate, informed by the robot's policy learning.
0:56

Data curation is crucial for post-training recipes. But how do we curate? Curation is usually manual & tedious. And, it's hard to tell if a strategy in the data will be reliable! We introduce an automatic way to curate, informed by the robot's policy learning.

Chelsea Finn

19,910 görüntüleme • 1 yıl önce

Excited to introduce 𝐋𝐀𝐏𝐀: the first unsupervised pretraining method for Vision-Language-Action models. Outperforms SOTA models trained with ground-truth actions 30x more efficient than conventional VLA pretraining 📝: 🧵 1/9
0:58

Excited to introduce 𝐋𝐀𝐏𝐀: the first unsupervised pretraining method for Vision-Language-Action models. Outperforms SOTA models trained with ground-truth actions 30x more efficient than conventional VLA pretraining 📝: 🧵 1/9

Joel Jang

46,018 görüntüleme • 1 yıl önce

🎧 What if you could own a piece of the next ChatGPT? @annakaslowskas joins Kenny Z of LayerZero to explain how your data could be worth more than a one-time payment. Listen in to the space below👇
0:32

🎧 What if you could own a piece of the next ChatGPT? @annakaslowskas joins Kenny Z of LayerZero to explain how your data could be worth more than a one-time payment. Listen in to the space below👇

vana

15,859 görüntüleme • 1 yıl önce

What would you tell your favorite anime character if you could speak to them? (via ZENSHU)
0:55

What would you tell your favorite anime character if you could speak to them? (via ZENSHU)

Crunchyroll

92,703 görüntüleme • 1 yıl önce

kylee this could be you if you listened to more nonwhite people music
0:35

kylee this could be you if you listened to more nonwhite people music

💋

33,210 görüntüleme • 5 ay önce

D2E Scaling Vision-Action Pretraining on Desktop Data for Transfer to Embodied AI
0:25

D2E Scaling Vision-Action Pretraining on Desktop Data for Transfer to Embodied AI

AK

25,117 görüntüleme • 9 ay önce

Concise, articulate, and easy to digest. - Less emphasis on shape and more on style influenced by the quality of players & what the opponents give you - Importance of organization for counterpressing situations via more passes in the oppositions half
0:38

Concise, articulate, and easy to digest. - Less emphasis on shape and more on style influenced by the quality of players & what the opponents give you - Importance of organization for counterpressing situations via more passes in the oppositions half

HP

398,772 görüntüleme • 2 yıl önce

“if you could be any animal what would you choose?”
1:29

“if you could be any animal what would you choose?”

Kevin Sanji

24,189 görüntüleme • 1 yıl önce

Larry Ellison openly describes a plan to unify all citizen data via digital ID into a single national database—designed to be fully “consumable” by AI models. Centralized data + AI = enormous power.
1:14

Larry Ellison openly describes a plan to unify all citizen data via digital ID into a single national database—designed to be fully “consumable” by AI models. Centralized data + AI = enormous power.

Truthseeker

11,770 görüntüleme • 5 ay önce

If you could be from another country what would it be
1:01

If you could be from another country what would it be

Mr Decent🦅

332,672 görüntüleme • 3 ay önce

If you could get HP to autograph ONE thing, what would it be? 😅
0:28

If you could get HP to autograph ONE thing, what would it be? 😅

Mumbai Indians

112,886 görüntüleme • 1 ay önce

here's Jeff Dean talking about how labs will do multi-epoch pretraining with heavy regularization to keep scaling even with limited data. no wonder slowrun gets so much attention from pretraining teams at big labs. pretraining is about to look very very different.
0:21

here's Jeff Dean talking about how labs will do multi-epoch pretraining with heavy regularization to keep scaling even with limited data. no wonder slowrun gets so much attention from pretraining teams at big labs. pretraining is about to look very very different.

Samip

127,068 görüntüleme • 3 ay önce

If you could give 1 piece of advice to men on how to be healthy, what would it be?
0:30

If you could give 1 piece of advice to men on how to be healthy, what would it be?

Bathmate

10,459 görüntüleme • 1 yıl önce

if you could recommend one song to someone getting into kpop what would it be?
2:20

if you could recommend one song to someone getting into kpop what would it be?

nugu archive 📁

51,991 görüntüleme • 4 ay önce

What if you could vote without leaving your house? Here's why Mobile Voting founder Bradley Tusk says creating a safe and secure way to vote via phone could be key to increasing voter turn out and reducing political extremism:
2:48

What if you could vote without leaving your house? Here's why Mobile Voting founder Bradley Tusk says creating a safe and secure way to vote via phone could be key to increasing voter turn out and reducing political extremism:

TED Talks

35,121 görüntüleme • 8 ay önce

What happens when vision🤝 robotics meet? Happy to share our new work on Pretraining Robotic Foundational Models!🔥 ARM4R is an Autoregressive Robotic Model that leverages low-level 4D Representations learned from human video data to yield a better robotic model. Berkeley AI Research😊
0:49

What happens when vision🤝 robotics meet? Happy to share our new work on Pretraining Robotic Foundational Models!🔥 ARM4R is an Autoregressive Robotic Model that leverages low-level 4D Representations learned from human video data to yield a better robotic model. Berkeley AI Research😊

Roei Herzig

62,711 görüntüleme • 1 yıl önce

Ever wondered what robots 🤖 could achieve if they could not just see – but also feel and hear? Introducing FuSe: a recipe for finetuning large vision-language-action (VLA) models with heterogeneous sensory data, such as vision, touch, sound, and more. Details in the thread 👇
0:49

Ever wondered what robots 🤖 could achieve if they could not just see – but also feel and hear? Introducing FuSe: a recipe for finetuning large vision-language-action (VLA) models with heterogeneous sensory data, such as vision, touch, sound, and more. Details in the thread 👇

Carlo Sferrazza

46,111 görüntüleme • 1 yıl önce

What is $REPPO? Reppo uses prediction markets to power AI data curation. Publishers submit data and domain experts stake $REPPO to assess quality within Datanets, programmable on-chain markets for any AI use case.
0:20

What is $REPPO? Reppo uses prediction markets to power AI data curation. Publishers submit data and domain experts stake $REPPO to assess quality within Datanets, programmable on-chain markets for any AI use case.

Kraken Listings

21,193 görüntüleme • 2 ay önce