Orion 14B πEnglish π¨π³δΈζ π―π΅ζ₯ζ¬θͺ π°π·νκ΅μ΄ π€ HuggingFace Mainpage π€ ModelScope Mainpage π¬ HuggingFace Demo π« ModelScope Demo πΊ GitHub π Tech Report Table of Contents π Model Introduction π Model Download π Model Benchmark π Model Inference π Declarations & License π₯ Company Introduction 1. Model Introduction Orion 14B series models are open source multilingual large language models trained from scratch by OrionStarAI. The base model is trained on 2.5T multilingual corpus, including Chinese, English, Japanese, Korean, etc, and it exhibits superior performance in these languages. For details, please refer to tech report. The Orion 14B series models exhibit the following features: Among models with 20B parameter scale level, Orion 14B Base model shows outstanding performance in comprehensive evaluations. Strong multilingual capabilities, significantly outperforming in Japanese and Korean testsets. The fine tuned models demonstrate strong adaptability, excelling in human annotated blind tests. The long chat version supports extremely long texts, performing exceptionally well at a token length of 200k and can support up to a maximum of 320k. The quantized versions reduce modβ¦
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy