# Huggingface model for inference in aws sagemaker

**URL:** <https://discuss.huggingface.co/t/huggingface-model-for-inference-in-aws-sagemaker/18485>\
**Category:** Amazon SageMaker\
**Created:** [May 29, 2022, 3:16pm UTC](https://discuss.huggingface.co/t/huggingface-model-for-inference-in-aws-sagemaker/18485 "2022-05-29T15:16:55Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![Jiiho](https://avatars.discourse-cdn.com/v4/letter/j/59ef9b/32.png) [@Jiiho](https://discuss.huggingface.co/u/Jiiho)\
**Post date:** [May 29, 2022, 3:16pm UTC](https://discuss.huggingface.co/t/huggingface-model-for-inference-in-aws-sagemaker/18485/1 "2022-05-29T15:16:55Z")

</div>

Hi:)  
I checked the model’s inference value through ‘predictor’ in ‘sagemaker’ as follows.

By the way, what does the ‘score’ in the 13th cell mean?

It is confusing because the value of the score is different according to the context and question.

Also, is there a way to display the decimal point of the ‘score’ to two digits?  
I’ve been searching for days but couldn’t find it. Any help would be greatly appreciated!

 ![image](https://us1.discourse-cdn.com/hellohellohello/original/2X/4/47eba9b354fa51ff360dc684dbd5e0d745731937.png)

---

<div class="post-metadata">

**Author:** ![philschmid](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/philschmid/32/38022_2.png) [@philschmid](https://discuss.huggingface.co/u/philschmid)\
**Post date:** [May 30, 2022, 6:36am UTC](https://discuss.huggingface.co/t/huggingface-model-for-inference-in-aws-sagemaker/18485/2 "2022-05-30T06:36:27Z")

</div>

Hello @Jiiho,

The `score` is the probability of the answer between 0-1. We are using the `question-answering` pipeline of transformers: [transformers/question\_answering.py at v4.19.2 · huggingface/transformers · GitHub](https://github.com/huggingface/transformers/blob/v4.19.2/src/transformers/pipelines/question_answering.py#L101)

You can modify this by creating your own `inference.py` and adjust the `post-processing` of the `pipeline`
