I’ve been trying to balance my data using the sample_from_datasets method but sometimes I get the following error :

WARNING:tensorflow:Your input ran out of data; interrupting training. Make sure that your dataset or generator can generate at least `steps_per_epoch * epochs` batches (in this case, 61800 batches). You may need to use the repeat() function when building your dataset

Is there a parameter I can set (with a callback) to make the fit function always run every epoch like the first epoch where it has Unknown amount of steps without having to do repeat and setting the steps_per_epoch?

I I’ve already tried setting the steps_per_epoch and repeating the dataset but I wish for it to run on all the data without have to repeat.