I never did the fine-tuning myself. It’s not that interesting to me. And I eventually lost interest in the leaderboard. It became increasingly clear that some submissions were training on the test set, and the whole thing was eventually shut down and rebooted. But I know the method is real, because I never used the leaderboard benchmarks for optimisation. The leaderboard was always just validation.
«Запасов газа осталось на два дня». Европа становится уязвимой из-за конфликта на Ближнем Востоке. Почему?00:54
，更多细节参见safew
I was able to find the issue and nudge it towards competitive speeds:
人 民 网 版 权 所 有 ，未 经 书 面 授 权 禁 止 使 用
/ Divide by the following number