How about this pruned model inference speed compared original model?
How about this pruned model inference speed compared original model?