# Visualizing attention heatmaps of layoutlmv3

**URL:** </static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fvisualizing-attention-heatmaps-of-layoutlmv3%2F32485%3E%5C
**Category:** 🤗Transformers\
**Created:** [February 25, 2023, 4:38am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fvisualizing-attention-heatmaps-of-layoutlmv3%2F32485 "2023-02-25T04:38:32Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![suryakml](https://avatars.discourse-cdn.com/v4/letter/s/f05b48/32.png) [@suryakml](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Fu%2Fsuryakml)%5C
**Post date:** [February 25, 2023, 4:38am UTC](/static-proxy?url=https%3A%2F%2Fdiscuss.huggingface.co%2Ft%2Fvisualizing-attention-heatmaps-of-layoutlmv3%2F32485%2F1 "2023-02-25T04:38:32Z")

</div>

I’m looking to interpret the attention outputs from the [`LayoutLMv3ForTokenClassification`](https://huggingface.co/docs/transformers/v4.26.1/en/model_doc/layoutlmv3#transformers.LayoutLMv3ForTokenClassification) model with sroie dataset from [`darentang/sroie`](https://huggingface.co/datasets/darentang/sroie) for the key information extraction task. I am trying to visualize the heatmaps with respect to the input text and image. Is it possible to do this with the `attentions` attribute of the model output?

Any help is appreciated! Thank you.