max iter parameter in shhh.flows


If I got a very small data set (so almost no error modeling), and I rise the max iter param, is the denoising will be better or will it make it worst ?
And if a got a very large dataset will it denoise it too much and “Damage” the sequences ?

Thanks in advance

If you have a small dataset, then you are unlikely to ever get to the max iter value. You might try changing mindelta from 1e-6 to something smaller, but I doubt it will help too much on a small dataset. I probably wouldn’t worry to much about it .

If i change the min delta to a lower value on a large dataset, is there a risk of damaging the sequences (with the error modeling being good in a large dataset ?)