How is the update for "train_on_batch" determined?

[gxlzj]

Dear developers and friends,

I am a new user to keras. My question is as fellows:

I see for "train_on_batch", the optimizer only do one single gradient update. But how is the learning rate/ momentum used in this setting?

For example, for SGD, we have a learning rate and momentum. When I use "train_on_batch" manually, will optimizer SGD automatically store the gradient update and use it as momentum when I use "train_on_batch" again next time?

[codekansas]

Yes, it will, consult here. Momentum is persistent between batches.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed after 30 days if no further activity occurs, but feel free to re-open a closed issue if needed.