GRPO (Group Relative Policy Optimization) - the core algorithm behind deepseek r1 - explained : https://t.co/qgRkPWWoTJ
— λux (@novasarc01) Jan 25, 2025
from Twitter https://twitter.com/novasarc01
January 25, 2025 at 05:27PM
via IFTTT
GRPO (Group Relative Policy Optimization) - the core algorithm behind deepseek r1 - explained : https://t.co/qgRkPWWoTJ
— λux (@novasarc01) Jan 25, 2025
No comments:
Post a Comment