How much data is roughly needed to train one's own dataset?
How much data is roughly needed to train one's own dataset?