Skip to main content
12 events
when toggle format what by license comment
S Apr 17, 2021 at 22:01 history suggested jameshfisher CC BY-SA 4.0
fix typo "plateau"
Apr 17, 2021 at 19:52 review Suggested edits
S Apr 17, 2021 at 22:01
Apr 6, 2021 at 10:04 comment added Ismael EL ATIFI The real question is : after steps_per_epoch batches have passed and the epoch is considered finished, will the next epoch start again from 1st batch of the dataset or will it continue from the next batch ? If it starts again at first batch, this is an issue as some batches in the dataset will never be used.
Feb 24, 2021 at 7:28 history edited pcko1 CC BY-SA 4.0
added example
Jun 1, 2020 at 17:38 comment added Corel @pcko1: What does it mean then, if steps_per_epoch is larger than num_samples // batch_size? Because I saw few cases like that.
Feb 17, 2020 at 22:24 comment added pcko1 Thank you for your comment. Do you have any references which show when get_updates is called? It would be interesting to see
Feb 17, 2020 at 22:09 comment added Bananach I didn't think so. You are saying steps_per_epoch allows for the learning rate to be adjusted during a pass through the samples, by setting steps_per_epoch < num_samples//batch_size. I am saying that that'll happen anyway even with steps_per_epoch=num_samples//batch_size, and it'll happen exactly steps_per_epoch often
Feb 17, 2020 at 21:54 comment added pcko1 @Bananach I believe we are saying the same thing, right?
Feb 17, 2020 at 18:03 comment added Bananach The learning rate is decayed at each step, not at each epoch, according to github.com/keras-team/keras/blob/…
Mar 16, 2019 at 14:29 comment added yamini goel I saw the documentation before posting the question, I couldn't understand what it means by num_samples. Does it mean total number of rows in x_train file?
Mar 16, 2019 at 13:07 history edited pcko1 CC BY-SA 4.0
added 27 characters in body
Mar 16, 2019 at 12:46 history answered pcko1 CC BY-SA 4.0