As part of our continued work to help assure the future security of deployed cryptographic systems, we recently released new code that will enable researchers to benchmark AI-based attacks — and compare them to new and existing attacks going forward
@aiatmeta
-
Meta releases Llama Guard 3 research for LLM safety advancement
By
–
Our hope in releasing this research openly is that it helps practitioners build better safeguard models and inspires further research in LLM safety. Llama Guard 3 1B paper https://
go.fb.me/5ofm27
Llama Guard 3 Vision paper https://
go.fb.me/2hitev -
Meta releases Llama Guard 3 safety research for multimodal AI
By
–
Following the release of our latest system level safeguards, today we're sharing new research papers outlining work on Llama Guard 3 1B & Llama Guard 3 Vision that support input/output safety on the edge and in multimodal prompts.
-
Llama Model Achieves 969 Tokens Per Second Inference Speed
By
–
969 tok/sec, impressive work from @cerebras
! We're excited to see what kinds of new applications developers can build with our most advanced Llama model at this inference speed! -
NVIDIA Optimizes Llama 3.2 Vision Models for Edge Devices
By
–
The NVIDIA team shared more on how they optimized Llama 3.2 on-device and vision models for performance and cost-efficiency from data-center scale all the way to low-power edge devices.
-
Startup Community Innovation with Llama Derivatives
By
–
It's energizing to see the new derivatives and innovation from the startup community building with Llama!
-
Meta Research Papers at EMNLP 2024 Conference
By
–
Whether you're at #EMNLP2024 in person or following from your feed, here are research papers being presented by AI research teams at Meta to add to your reading list. Distilling System 2 into System 1: https://
go.fb.me/5l9832 Altogether: Image Captioning via Re-aligning -
NeuralFeels: Neural Fields for Visuotactile In-Hand Manipulation
By
–
Newly published in this issue of Science Robotics today from Meta FAIR: NeuralFeels with neural fields — Visuotactile perception for in-hand manipulation https://
go.fb.me/2cv7r6 -
Next Hackathon Kicks Off Tonight Online and in San Francisco
By
–
Interested in joining the next hackathon? We're kicking off tonight, online and in San Francisco with @lablabai!https://t.co/FVNHz1jdiZ
— AI at Meta (@AIatMeta) 8 novembre 2024Interested in joining the next hackathon? We're kicking off tonight, online and in San Francisco with @lablabai
! -
First Llama Hackathon in India: 270+ Developers, 75+ Projects
By
–
Together w/ @Reskill
, we hosted the first Llama Hackathon in India bringing together 270+ developers & 25+ mentors from across industries creating in 75+ projects in just 30h of hacking! See the recap and details on some of the top projects here https://
go.fb.me/nmsa9c