From 012afab40ce0762b1e5cf2689293f78ac8d09746 Mon Sep 17 00:00:00 2001 From: Rutam Prita Mishra Date: Tue, 3 Oct 2023 04:11:04 +0530 Subject: [PATCH] Created magorshunov/layoutlm-invoices README.md File --- magorshunov-layoutlm-invoices/README.md | 37 +++++++++++++++++++++++++ 1 file changed, 37 insertions(+) create mode 100644 magorshunov-layoutlm-invoices/README.md diff --git a/magorshunov-layoutlm-invoices/README.md b/magorshunov-layoutlm-invoices/README.md new file mode 100644 index 0000000..5239625 --- /dev/null +++ b/magorshunov-layoutlm-invoices/README.md @@ -0,0 +1,37 @@ +![Cover](https://user-images.githubusercontent.com/66431403/267597217-2285216a-209c-466b-a0da-ab610356c2af.png) + +# layoutlm-invoices + +## DagsHub Repository: https://dagshub.com/Rutam21/layoutlm-invoices + +## Source: [HuggingFace layoutlm-invoices Model](https://huggingface.co/magorshunov/layoutlm-invoices) + +# Description + +This is a fine-tuned version of the multi-modal LayoutLM model for the task of question answering on invoices and other documents. It has been fine-tuned on a proprietary dataset of invoices as well as both SQuAD2.0 and DocVQA for general comprehension. + +# Fine Tune Results + +## Non-consecutive tokens + +Unlike other QA models, which can only extract consecutive tokens (because they predict the start and end of a sequence), this model can predict longer-range, non-consecutive sequences with an additional classifier head. For example, QA models often encounter this failure mode: + +### Before + +![Before Results](https://dagshub.com/Rutam21/layoutlm-invoices/raw/main/before.png) + +### After + +However this model is able to predict non-consecutive tokens and therefore the address correctly. + +![After Results](https://dagshub.com/Rutam21/layoutlm-invoices/raw/main/after.png) + +# License + +This model is available on HuggingFace under the CC by NC-SA 4.0 License. + +# Citation + +```citation +This model was created by the team at Impira. +```