Seonglae ChoMobile AI, Edge AI
Edge AI Tools
Mobile LLM below 1b (or SmolLM2, LLaMa-1b, Qwen 1b, Hymba )
OmniVision-968M: World's Smallest Vision Language Model
Pocket-size multimodal model with 9x token reduction for on-device deployment
https://nexa.ai/blogs/omni-vision
MobileLLM - a facebook Collection
Optimizing Sub-billion Parameter Language Models for On-Device Use Cases (ICML 2024) https://arxiv.org/abs/2402.14905
https://huggingface.co/collections/facebook/mobilellm-6722be18cb86c20ebe113e95
HuggingFaceTB/SmolLM2-1.7B-Instruct · Hugging Face
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
https://huggingface.co/HuggingFaceTB/SmolLM2-1.7B-Instruct
vision model (256M 500M 2.2B)
ds4sd/SmolDocling-256M-preview · Hugging Face
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
https://huggingface.co/ds4sd/SmolDocling-256M-preview
SmolVLM - small yet mighty Vision Language Model
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
https://huggingface.co/blog/smolvlm
SmolVLM Grows Smaller – Introducing the 256M & 500M Models!
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
https://huggingface.co/blog/smolervlm
dataset SmolTalk
HuggingFaceTB/smoltalk · Datasets at Hugging Face
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
https://huggingface.co/datasets/HuggingFaceTB/smoltalk
Solution is LoRA like adaptation technique?
Introducing Apple’s On-Device and Server Foundation Models
At the 2024 Worldwide Developers Conference, we introduced Apple Intelligence, a personal intelligence system integrated deeply into…
https://machinelearning.apple.com/research/introducing-apple-foundation-models
Meta and Qualcomm team up to run big A.I. models on phones
Large language models (LLMs) are the technology that underpin applications like OpenAI's ChatGPT that can return out text that resembles human output.
https://www.cnbc.com/2023/07/18/meta-and-qualcomm-team-up-to-run-big-ai-models-on-phones.html
Cognitive core could be extremely small
No Priors Ep. 80 | With Andrej Karpathy from OpenAI and Tesla
Andrej Karpathy joins Sarah and Elad in this week of No Priors. Andrej, who was a founding team member of OpenAI and the former Tesla Autopilot leader, needs no introduction. In this episode, Andrej discusses the evolution of self-driving cars, comparing Tesla's and Waymo’s approaches, and the technical challenges ahead. They also cover Tesla’s Optimus humanoid robot, the bottlenecks of AI development today, and how AI capabilities could be further integrated with human cognition. Andrej shares more about his new mission Eureka Labs and his insights into AI-driven education and what young people should study to prepare for the reality ahead.
Sign up for new podcasts every week. Email feedback to show@no-priors.com
Follow us on Twitter: @NoPriorsPod | @Saranormous | @EladGil | @Karpathy
Show Notes:
0:00 Introduction
0:33 Evolution of self-driving cars
2:23 The Tesla vs. Waymo approach to self-driving
6:32 Training Optimus with automotive models
10:26 Reasoning behind the humanoid form factor
13:22 Existing challenges in robotics
16:12 Bottlenecks of AI progress
20:27 Parallels between human cognition and AI models
22:12 Merging human cognition with AI capabilities
27:10 Building high performance small models
30:33 Andrej’s current work in AI-enabled education
36:17 How AI-driven education reshapes knowledge networks and status
41:26 Eureka Labs
42:25 What young people study to prepare for the future
https://www.youtube.com/watch?v=hM_h0UA7upI
