Afshin Dehghan
@afshin_dn
Followers
72
Following
0
Media
6
Statuses
10
Joined December 2023
When training LLMs, dataset size & quality matter as much as architecture. Scaling laws show: 📈 More compute → broader, less filtered data 📷 Less compute → tighter more curated datasets Small models need precision. Big models thrive on diversity. Optimize accordingly.
0
0
12
Yesterday we shared our latest work on pretraining data curation. What if we stop guessing which data is “good” and directly match pretraining data to the benchmarks we care about? 📄 https://t.co/Mvea0rJ8vc
#AIResearch #llm #DataCuration #Pretraining #ScalingLaws
0
4
23
Excited to share our new work: “Language Models Improve When Pretraining Data Matches Target Tasks” Yes, it sounds obvious (and it is!), but typically this only happens implicitly and indirectly: intuitively select data → benchmark → refine → repeat. We wondered: what
7
49
408
Incredibly proud of the work across teams in delivering the latest version of Visual Intelligence. Visual Intelligence makes it faster to do more with what’s right in front of you. #WWDC25 #visualintelligence #AppleIntelligence
0
0
1
Very excited to announce our final line-up of fantastic speakers at this year's @CVPR workshop on Open-World 3D Scene Understanding with Foundation Models ✨ #OpenSUN3D #cvpr2025 📆 June 12, 2pm-6pm 🏡 https://t.co/XqA2dyAp2Q
1
7
72
Excited to share that we have recently released the source code for FlexTok, bringing a fresh perspective to tokenization. Code on GitHub: https://t.co/ApWNbE2ZO6. Project Page: https://t.co/MlDKYAfSLz
#FlexTok #Tokenization #MachineLearning #MLResearch #OpenSource #AI
0
7
37
🚀 Model and data for our CubifyAnything project are now released! 🔗 https://t.co/d0VoQUaa0A
#SpatialReasoning #3DObjectDetection #transformers #detection #ai #genai
0
1
4
We'll present at NeurIPS, today at 5pm CST. Spotlight #1022. Effectively bringing sensory modalities to large models is one way to make them more grounded, and ultimately have a more complete World Model. This is a step in that direction hopefully, and more will come.
4M exhibits having learned a solid cross-modal representation. We can use the various modalities to probe how 4M reconciles unusual inputs by manipulating one part of it while keeping the remainder fixed. (8/n)
1
9
71
We are releasing the 1st version of 4M, a framework for training multimodal foundation models across tens of modalities & tasks, based on scalable masked modeling. Joint effort by @EPFL_en & @Apple. 4M: Massively Multimodal Masked Modeling 🌐 https://t.co/usE17pnXf9 🧵1/n
8
132
603