* Pruning for GPT attn heads * The code formatted according to the transformers requirements * Update run_prune_gpt.py * Update run_prune_gpt.py