GPT-2 124M trained from scratch – 2.48B tokens, 2.8h on 4 A40s

2026-10-15来源 Hacker News NewEnglish

Article URL: https://github.com/upul/Build-GPT-2

Comments URL: https://news.ycombinator.com/item?id=50104656

Points: 1

# Comments: 0

原文：https://www.a8q9.cn/en/article/gpt-2-124m-trained-from-scratch-2-48b-tokens-2-8h-on-4-a40s-5a4205.html

相关文章