I have a question regarding the preprocessing of the raw signal data. I was wondering if the raw signal data is just scaled or if you perform some form of additional normalization (per read) before model training / inference.
The provided k-mer models are given as expected values (normalized), so I wanted to know if the basecalling process was done in a similar manner.
Hi,
I have a question regarding the preprocessing of the raw signal data. I was wondering if the raw signal data is just scaled or if you perform some form of additional normalization (per read) before model training / inference. The provided k-mer models are given as expected values (normalized), so I wanted to know if the basecalling process was done in a similar manner.