# Eval Loss spike Seq2seq Trainer Resume from Checkpoint

**URL:** <https://discuss.huggingface.co/t/eval-loss-spike-seq2seq-trainer-resume-from-checkpoint/7067>\
**Category:** 🤗Transformers\
**Created:** [June 22, 2021, 2:42pm UTC](https://discuss.huggingface.co/t/eval-loss-spike-seq2seq-trainer-resume-from-checkpoint/7067 "2021-06-22T14:42:47Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![MillyXXX](https://avatars.discourse-cdn.com/v4/letter/m/a9a28c/32.png) [@MillyXXX](https://discuss.huggingface.co/u/MillyXXX)\
**Post date:** [June 22, 2021, 2:42pm UTC](https://discuss.huggingface.co/t/eval-loss-spike-seq2seq-trainer-resume-from-checkpoint/7067/1 "2021-06-22T14:42:47Z")

</div>

I ran Trainer.train(resume\_from\_checkpoint = True). The eval loss spiked from 1.x to 5.x, but training loss is decreasing consistently, any possible reasons for this? Thanks

 ![image](https://us1.discourse-cdn.com/hellohellohello/original/2X/1/1687b4b79a26abd931c881ce31c40dca166c7528.jpeg)