JEPA is just LoRA in disguise. New Insights.
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Unlock all features
FREE: Get instant access to 10 AI summaries, chats, or transcripts per day.
Related videos
Fox News JUST Aired THIS DISTURBING Trump Photo!
Luke Beasley
33.2k views
I just died inside…
Adam Mockler
305.5k views
LIVE | Karen Read Discovery Fights over Colin Albert Deposition. Is the Judge over it?
Emily D. Baker
82.3k views
Mom Discovers Her 13 Year Old Son Just Killed Someone
Dr Insanity
2.0m views
THIS IS HOW DEMOCRATS JUST LOST
Timcast
82.1k views
LIVE COURT | Karen Read Civil cases back in court for discovery fights.
Emily D. Baker
73.3k views
NEW GOP MAP IS IN, Democrats May LOSE The MIDTERMS | Timcast IRL
Timcast IRL
100.9k views
“This Could Destroy the Church” NEW Lost Ancient Text Discovered | Matt LaCroix
Danny Jones
13.0k views
Epstein DISASTER Uncovered in NEW DROP
Adam Mockler
188.3k views
The New Cause of Insulin Resistance is NOT Just Sugar
Thomas DeLauer
17.2k views
Top Comments (10)
Me encanta que te bases en papers científicos y novedosos. Saludos. Sigue asi
Maybe joint embedding isn't the way, but I'm still hyped for bong embedding and edibles embedding.
27:08 So it seems that JEPA may still have issues handling out-of-distribution data.
My mental model: it's a parser, a feature extractor
So it's not the AGI?😢
Well grounded presentation.
Im building datasetless ai🤓
The question is one of autoformalization: how to construct automatically something useful out of trillion-scale datasets rather than doing it ourselves
❤😂 científico se tardan mucho en llegar a este nivel de comprensión❤
This exactly what I was thinking especially when I was listening to yan lecunn’s presentations and so on
Unlock the Data Inside
Turn Videos into Knowledge
- Get FREE 10/day: transcripts, summaries, chats
- Chat with videos, export text & PDF
- $1 free API credit for RAG, chatbots & research
Free forever plan • All features unlocked
Top Comments (10)
Me encanta que te bases en papers científicos y novedosos. Saludos. Sigue asi
Maybe joint embedding isn't the way, but I'm still hyped for bong embedding and edibles embedding.
27:08 So it seems that JEPA may still have issues handling out-of-distribution data.
My mental model: it's a parser, a feature extractor
So it's not the AGI?😢
Well grounded presentation.
Im building datasetless ai🤓
The question is one of autoformalization: how to construct automatically something useful out of trillion-scale datasets rather than doing it ourselves
❤😂 científico se tardan mucho en llegar a este nivel de comprensión❤
This exactly what I was thinking especially when I was listening to yan lecunn’s presentations and so on