Sources
Loading...
Additional media
Loading...
DeepNewz, mobile.
People-sourced. AI-powered. Unbiased News.
MBZUAI and LLama3 teams have released new large language models (LLMs) named K2-65B and LLama3 8B Vision, respectively. K2-65B, developed in collaboration with PetuumInc and llm360, is a cost-effective open-source LLM with 65 billion parameters, outperforming LLama 2 70B. LLama3 8B Vision, close to GPT4V & GPT4o, features architecture changes like SIGLIP vs CLIP and a strong projection layer from SIGLIP to the LLama3 embedding.
Ready to explore our latest #LLM? K2-65B is a fully-reproducible large language model that outperforms Llama 2 70B and uses 35% less computing. Go straight to the source with all open-sourced artifacts, including code, data, model checkpoints, intermediate results, and more:…
Check out #K2, a fully-open 65B LLM released by @llm360 Matching the performance of #Llama2 70B, #K2 is among the most powerful LLMs made fully transparent! Over the past 6 months, @llm360 has made open a series of LLMs across different tiers, all with open weights,… https://t.co/ITsQqSZUpO
MAP-Neo: Highly Capable and Transparent Bilingual Large Language Model Series - Truly opensourced (e.g. training script, data, checkpoints available) - Up to 7B params trained on 4.5T tokens proj: https://t.co/RdWgOvhJzq abs: https://t.co/qF7xFZCZip https://t.co/znbFznIu9L
