Apple Intelligence Foundation Language Models

Apple Intelligence Foundation Language Models

We present foundation language models developed to power Apple Intelligence features, including a ~3 billion parameter model designed to run efficiently on devices and a large server-based language model designed for Private Cloud Compute. These models are designed to perform a wide range of tasks e...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:arXiv.org 2024-07
Hauptverfasser: Gunter, Tom, Wang, Zirui, Pang, Ruoming, Bowen, Zhang, Chung-Cheng, Chiu, Qiu, David, Yin, Dong, Yin, Guoli, Huang, Haoshuo, Wang, Jianyu, Ye, Ke, Chen, Qibin, Keunebroek, Quentin, Wiseman, Sam, Evans, Syd, Tao, Lei, Rathod, Vivek, Kong, Xiang, Du, Xianzhi, Li, Yanghao, Wang, Yongqiang, Rashid, Al, Albin, Madappally Jose, Bencomo, Alfredo, Vanderby, Allison, Hansen, Andrew, Jain, Ankur, Wu, Bugu, Maalouf, Charlie, Erdenebileg, Chinguun, Dulhanty, Chris, Moritz, Dominik, Kang, Doug, Ladd, Evan, Shi, Fangping, Bai, Felix, Chu, Frank, Kotek, Hadas, Hannah Gillis Coleman, Li, Jane, Bigham, Jeffrey, Lai, Jeff, Cheung, Jessica, Jiulong Shan, Zhou, Joe, Li, John, Qin, Jun, Singh, Karanjeet, Cordell, Maria, Hay, Nicole, Shahdadpuri, Nilesh, Godwin, Otto, Dighe, Pranay, Shah, Sanskruti, Guha, Saptarshi, Sirovica, Sasha, Ma, Shuang, Kim, Sulgi, Jayaram, Suma, Paidi, Varsha, Kumar, Vivek, Wang, Xin, Zheng, Xin, Walker, Cheng, Shrager, Yael, Yang, Ye, Tanaka, Yasu, Guo, Yihao, Meng, Yunsong, Zhao Tang Luo, Aygar, Alp, Walkingshaw, Andrew, Lin, Antonie, Ramerth, Brent, Bartels, Chris, Chaney, Chris, Hochstrasser, Gabriel, Seguin, Guillaume, Belousova, Irina, Pelemans, Joris, Yang, Karen, Najibi, Mahyar, Horton, Max, Cho, Minsik, Bhendawade, Nikhil, Dong, Patrick, Maj, Piotr, Agrawal, Pulkit, Fu, Qichen, Poston, Regan, Xu, Sam, Liu, Shuangning, Heeramun, Tashweena, Merth, Thomas, Rayala, Uday, Cui, Victor, Vivek Rangarajan Sridhar, Zhang, Wencong, Zhao, Yang, Yin Xia
Format: Artikel
Sprache:eng
Schlagworte:
Intelligence
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:We present foundation language models developed to power Apple Intelligence features, including a ~3 billion parameter model designed to run efficiently on devices and a large server-based language model designed for Private Cloud Compute. These models are designed to perform a wide range of tasks efficiently, accurately, and responsibly. This report describes the model architecture, the data used to train the model, the training process, how the models are optimized for inference, and the evaluation results. We highlight our focus on Responsible AI and how the principles are applied throughout the model development.
ISSN:2331-8422