Efficient Mathematical Reasoning Models via Dynamic Pruning and Knowledge Distillation

Open in new window