Currently, the torch and transformer (ML) dependencies are required (loaded) by the core. Try to avoid this with lazy loading.
Currently, the torch and transformer (ML) dependencies are required (loaded) by the core.
Try to avoid this with lazy loading.