GLiNER
GLiNER copied to clipboard
Question about max_types
Excellent job!
I'd like to inquire about the 'max_types' setting. Suppose 'max_types' is configured to a limit of 25; does this imply that we cannot employ more than 25 types at the time of inference? My concern is that this limitation might affect the model's performance, as inputs beyond this range could be considered out-of-domain. What would be the recommended approach to address this issue?
Should we divide the types into smaller groups, each with less than 25 types, perform multiple inferences, and then select the type with the highest score? Or should we increase the 'max_types' threshold? However, I wonder if this might complicate the training process for the model.