# Runseq2seq\_qa.py not storing predictions

**URL:** <https://discuss.huggingface.co/t/runseq2seq-qa-py-not-storing-predictions/30787>\
**Category:** 🤗Transformers\
**Created:** [January 31, 2023, 4:02am UTC](https://discuss.huggingface.co/t/runseq2seq-qa-py-not-storing-predictions/30787 "2023-01-31T04:02:50Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![gokul427](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/gokul427/32/13281_2.png) [@gokul427](https://discuss.huggingface.co/u/gokul427)\
**Post date:** [January 31, 2023, 4:02am UTC](https://discuss.huggingface.co/t/runseq2seq-qa-py-not-storing-predictions/30787/1 "2023-01-31T04:02:50Z")

</div>

Hi,

Im using runseq2seq\_qa.py with t5 model for QA task, but while prediction its not producing f1 metrics and not storing the output. Please suggest.

EvalLoopOutput(predictions=None, label\_ids=None, metrics={‘test\_loss’: 3.253300666809082, ‘test\_runtime’: 30.7896, ‘test\_samples\_per\_second’: 25.138, ‘test\_steps\_per\_second’: 3.15}, num\_samples=774)

cmd used: python run\_seq2seq\_qa\_own.py --model\_name\_or\_path /tmp/t5\_small\_squad\_own --dataset\_loading\_script load\_sqlike\_dataset\_both.py --context\_column context --question\_column question --answer\_column answers --do\_predict --per\_device\_train\_batch\_size 8 --learning\_rate 3e-5 --num\_train\_epochs 2 --max\_seq\_length 384 --doc\_stride 128 --output\_dir /tmp/t5\_small\_squad\_own/