Unable to reproduce the results for param_first and param_second in the paper after finetuning.

horseee / LLM-Pruner

[NeurIPS 2023] LLM-Pruner: On the Structural Pruning of Large Language Models. Support Llama-3/3.1, Llama-2, LLaMA, BLOOM, Vicuna, Baichuan, TinyLlama, etc.

https://arxiv.org/abs/2305.11627

Apache License 2.0

836 stars 98 forks source link

Unable to reproduce the results for param_first and param_second in the paper after finetuning. #52

Open danyal97 opened 7 months ago

danyal97 commented 7 months ago

Could you please tell me which command you used for post training of model to get results for element 1 and element 2 method.