mirror of
https://github.com/lllyasviel/ControlNet.git
synced 2026-04-24 03:00:54 -04:00
training
This commit is contained in:
@@ -259,7 +259,7 @@ Also, if your dataset is large, you may want to end the training with a few thou
|
||||
|
||||
Also, if you unlock some original layers, you may want a lower learning rate, like 2e-6.
|
||||
|
||||
## Other Considerations: the sudden converge phenomenon and gradient accumulation
|
||||
## More Consideration: Sudden Converge Phenomenon and Gradient Accumulation
|
||||
|
||||

|
||||
|
||||
|
||||
Reference in New Issue
Block a user