Update: Yandex/AliceAI 80B-A3B fine tune progress

loss curve (taken from the last micro of every step, to explain the variation) some help from gemini 3.8 flash high About 40% of the way done with the initial fine tune. The loss is so spiky because I accidentally used the last loss of each micro, rather than the average of each step The train

中文全文翻译正在补齐，可先查看下方英文原文。

loss curve (taken from the last micro of every step, to explain the variation) some help from gemini 3.8 flash high About 40% of the way done with the initial fine tune. The loss is so spiky because I accidentally used the last loss of each micro, rather than the average of each step The train