Stay up to date on all Petuum news and content
CrystalCoder
7B code and text model trained on 1.4T tokens combing the best of StarCoder and Llama.
Amber
7B English model trained on 1.2T tokens and released with 360 checkpoints fostering R&D.
K2
65B English model charting a blueprint towards Artificial General Intelligence (AGI).
Adapt best-in-class models on domain data
Easily learns different tasks with few examples
Integrates with your tools and processes
Stay up to date on all Petuum news and content