Dear cryosparc community,
I am working with a heterogenous sample and I cannot clean up beyond a point with 2D classification. My ab-initio reconstructions with default settings was noisy and did not yield anything meaningful.
I increased the initial/final batch size to a very large number say 30,000 and 100,000 particles. Although it takes about a week to finish, I start to see some meaningful ab-initio models.
Is using such a high batch size for ab-initio reconstruction a valid approach?