The introduction of StarCoder2, the latest code-generating AI, marks a significant advancement in the field of programming. Developed with a 16k token context and trained on over 4 trillion tokens from the Stack v2, the largest code dataset with 900B+ tokens, StarCoder2 has been designed to outperform its predecessor, StarCoder1, significantly. Available in three sizes, it supports more than 600 programming languages. Notably, StarCoder2 includes models like the 15B version, which surpasses the performance of CodeLlama 34B, offering a 16,384 context window. Furthermore, StarCoder2 is optimized for performance and cost, capable of matching CodeLlama 33B in code completion benchmarks at twice the speed and half the cost. This groundbreaking AI runs on most GPUs and is fully open, including all code, data, and models. Collaborations between ServiceNow, Hugging Face, and Nvidia have been instrumental in launching StarCoder2, aiming to facilitate the development of enterprise applications using Generative AI. The release also includes smol-StarCoder2 models at 3B and 7B sizes.
StarCoder2 is here!💫 A family of open LLMs enabling users with powerful performance and cost optimization. StarCoder 15B matches CodeLlama 33B in code completion benchmarks at 2x speed and 2x as cheap to train and use in production.🤯 https://t.co/zGUBDJi4IB https://t.co/DMb4itWxPD
StarCoder 2 is a code-generating AI that runs on most GPUs https://t.co/C4AXmj0Dar
StarCoder2 is the new SOTA code completion model! https://t.co/CQIbF2PHMz https://t.co/5x4PzAV9yr
Introducing StarCoder2 15B 🌟 > Beats CodeLlama 34B. > 16,384 context window. > Trained in 600+ programming languages from The Stack v2. > Trained on Fill-in-the-middle objective on 4 trillion + tokens. Along with that, we release smol-StarCoder2 3B & 7B ⭐ > 16K context… https://t.co/Sgi6eCK4Rv
Introducing StarCoder 2 ⭐️ The most complete open Code-LLM 🤖 StarCoder 2 is the next iteration for StarCoder and comes in 3 sizes, trained 600+ programming languages on over 4 Trillion tokens on Stack v2. It outperforms StarCoder 1 by margin and has the best overall performance… https://t.co/LVclRcq5ZM
We're live with StarCoder2! https://t.co/nyS7YPszYc
StarCoder 2 is a code-generating AI that runs on most GPUs: https://t.co/c4JLvwpyJE by TechCrunch #infosec #cybersecurity #technology #news
Introducing: StarCoder2 and The Stack v2 ⭐️ StarCoder2 is trained with a 16k token context and repo-level information for 4T+ tokens. All built on The Stack v2 - the largest code dataset with 900B+ tokens. All code, data and models are fully open! https://t.co/fM7GinxJBd https://t.co/NUeRjHEa05
.@ServiceNow + @huggingface + @nvidia = 🚀 Together, we’ve teamed up to launch StarCoder2: a family of open-access LLMs to help developers use GenAI to build enterprise applications. https://t.co/JzmKBrC42e #PutAIToWork https://t.co/a3GnyJA8PP