Fugaku-LLM: Japan's Supercomputer Trains Revolutionary Language Model for AI in Science and Social S

Опубликовано: 14 Май 2024
на канале: AI Insight News
66
0

SUBSCRIBE CHANNEL: https://bit.ly/AIInsightNews
-----------------
The Fugaku-LLM, a large language model developed in Japan using the Fugaku supercomputer, has 13 billion parameters and outperforms other models in tasks related to humanities and social sciences. It was trained on a combination of Japanese, English, mathematics, and code data. The model's results are available on GitHub and Hugging Face for further development, with potential applications in AI for science and social simulations. The project was a collaboration between multiple institutions. Comments on the post discuss the use of CPUs instead of GPUs for training, comparisons to other models like GPT-4, the efficiency of the training process, and the challenges of decentralized architecture for training models.

🔗 https://www.fujitsu.com/global/about/...

#AI #Language Model #GPT #GPT4 #LLM #OpenAI


Смотрите видео Fugaku-LLM: Japan's Supercomputer Trains Revolutionary Language Model for AI in Science and Social S онлайн без регистрации, длительностью часов минут секунд в хорошем качестве. Это видео добавил пользователь AI Insight News 14 Май 2024, не забудьте поделиться им ссылкой с друзьями и знакомыми, на нашем сайте его посмотрели 6 раз и оно понравилось людям.