# Accuracy drops using Gradient checkpointing

**URL:** <https://discuss.huggingface.co/t/accuracy-drops-using-gradient-checkpointing/53953>\
**Category:** 🤗Transformers\
**Created:** [September 7, 2023, 2:22pm UTC](https://discuss.huggingface.co/t/accuracy-drops-using-gradient-checkpointing/53953 "2023-09-07T14:22:33Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![jaideepcs](https://avatars.discourse-cdn.com/v4/letter/j/ecae2f/32.png) [@jaideepcs](https://discuss.huggingface.co/u/jaideepcs)\
**Post date:** [September 7, 2023, 2:22pm UTC](https://discuss.huggingface.co/t/accuracy-drops-using-gradient-checkpointing/53953/1 "2023-09-07T14:22:33Z")

</div>

Hi,  
I am not able to replicate same performance as without using gradient checkpointing.  
With Gradient ckp eval metrics is half without gradient ckp. I use lora adpaters . Am i missing any thing ?