增量训练问题 #2823
CyanMystery
started this conversation in
Ideas
增量训练问题
#2823
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
作者你好！ 我发现增量训练时 使用后面数据集训练的模型在web推理 前面训练数据的推理不准确， 如果是两个数据集一次训练效果还是挺不错的，请问这个有办法解决吗，总不能随着后续数据的增加 又要重头训练一遍吧？那开销也太大了
Beta Was this translation helpful? Give feedback.
All reactions