Hi Ravers,
I would like to train on a very small oboe-multiphonics-dataset (ca. 1h of sound - I know that is technically too small, but I’m rather interested rather in the unexpected results it might give).
Since it’s a very small set: what would be the best configuration for the training and how many epochs should it more or less train?
Thanks in advance for the tips!
All the best,
Joost