Training a model as large and capable as Llama 3.1 405B was no simple task. The model was trained on over 15 trillion tokens over the course of several months requiring over 16K @NVIDIA H100 GPUs — making it the first Llama model ever trained at this scale. We also used the 405B
@aiatmeta
-
Meta Releases Llama 3.1: New Open Source AI Models Including 405B
By
–
Starting today, open source is leading the way. Introducing Llama 3.1: Our most capable models yet. Today we’re releasing a collection of new Llama 3.1 models including our long awaited 405B. These models deliver improved reasoning capabilities, a larger 128K token context
-
JASCO: Music Generation Model with Improved Controllability
By
–
We recently announced JASCO, a music-generation model with improved controllability using conditioning inputs like chords or beat. We've shared more details on this work in the research paper https://
go.fb.me/sou2ox -
Happy Birthday Ollama: Celebrating Community Contributions
By
–
What a year! Happy birthday Ollama, and thank you for everything you do for the Llama community!
-
Llama 3 Model Size Increase: 7B to 8B Explanation
By
–
Some context on why our smallest Llama 3 model went from 7B → 8B. More details on the changes to the tokenizer in the full conversation with @astongzhangAZ ➡️ https://t.co/cKuUfwmuZ4 pic.twitter.com/OfqCJtfSJW
— AI at Meta (@AIatMeta) 16 juillet 2024Some context on why our smallest Llama 3 model went from 7B → 8B. More details on the changes to the tokenizer in the full conversation with @astongzhangAZ https://
youtu.be/Tmdk_H2WDj4 -
Llama 2 Project Building: CoLab Notebook from MetaConnect2023
By
–
P.S. Want to start building your own project with Llama 2? Here's the full CoLab notebook from our workshop at #MetaConnect2023 today https://
bit.ly/3RDEX7K -
Llama 2 Ecosystem: Past, Present and Future Updates
By
–
There's a lot to talk about with Llama 2 at #MetaConnect2023. We've published a new post on our blog on the past, present and future of the Llama ecosystem covering some of the updates we shared at Connect and more
-
MetaConnect 2023 Day Two: Lightning Talks on Llama and AI Products
By
–
Ready for day two of #MetaConnect2023? Join us for lightning talks on the Llama ecosystem, AI powering our next-generation product experiences & more starting at 11:15am PT! See the full schedule: http://
metaconnect.com Watch the lightning talks: https://
facebook.com/metafordevelop
ers
… -
Segment Anything Model Powers Upcoming Instagram Backdrop Feature
By
–
Segment Anything https://
bit.ly/3PBOmtU This popular model released earlier this year is an important component of the upcoming Backdrop feature in Instagram.