From 723ec3f9f85621d22c2a9fe570d3b04a2d773243 Mon Sep 17 00:00:00 2001
From: Ankur Goyal <ankrgyl@gmail.com>
Date: Wed, 24 Aug 2022 17:18:56 +0000
Subject: [PATCH] Upload TFLayoutLMForQuestionAnswering

---
 README.md   | 69 +++++++++++++++++++++++------------------------------
 config.json |  8 ++++---
 tf_model.h5 |  3 +++
 3 files changed, 38 insertions(+), 42 deletions(-)
 create mode 100644 tf_model.h5

diff --git a/README.md b/README.md
index f09c45c..d9aec21 100644
--- a/README.md
+++ b/README.md
@@ -1,57 +1,48 @@
 ---
-language: en
-thumbnail: https://uploads-ssl.webflow.com/5e3898dff507782a6580d710/614a23fcd8d4f7434c765ab9_logo.png
 license: mit
+tags:
+- generated_from_keras_callback
+model-index:
+- name: layoutlm-document-qa
+  results: []
 ---
 
-# LayoutLM for Visual Question Answering
+<!-- This model card has been generated automatically according to the information Keras had access to. You should
+probably proofread and complete it, then remove this comment. -->
 
-This is a fine-tuned version of the multi-modal [LayoutLM](https://aka.ms/layoutlm) model for the task of question answering on documents. It has been fine-tuned on
+# layoutlm-document-qa
 
-## Model details
+This model is a fine-tuned version of [impira/layoutlm-document-qa](https://huggingface.co/impira/layoutlm-document-qa) on an unknown dataset.
+It achieves the following results on the evaluation set:
 
-The LayoutLM model was developed at Microsoft ([paper](https://arxiv.org/abs/1912.13318)) as a general purpose tool for understanding documents. This model is a fine-tuned checkpoint of [LayoutLM-Base-Cased](https://huggingface.co/microsoft/layoutlm-base-uncased), using both the [SQuAD2.0](https://huggingface.co/datasets/squad_v2) and [DocVQA](https://www.docvqa.org/) datasets.
 
-## Getting started with the model
+## Model description
 
-To run these examples, you must have [PIL](https://pillow.readthedocs.io/en/stable/installation.html), [pytesseract](https://pypi.org/project/pytesseract/), and [PyTorch](https://pytorch.org/get-started/locally/) installed in addition to [transformers](https://huggingface.co/docs/transformers/index).
+More information needed
 
-```python
-from transformers import AutoTokenizer, pipeline
+## Intended uses & limitations
 
-tokenizer = AutoTokenizer.from_pretrained(
-    "impira/layoutlm-document-qa",
-    add_prefix_space=True,
-    trust_remote_code=True,
-)
+More information needed
 
-nlp = pipeline(
-    model="impira/layoutlm-document-qa",
-    tokenizer=tokenizer,
-    trust_remote_code=True,
-)
+## Training and evaluation data
 
-nlp(
-    "https://templates.invoicehome.com/invoice-template-us-neat-750px.png",
-    "What is the invoice number?"
-)
-# {'score': 0.9943977, 'answer': 'us-001', 'start': 15, 'end': 15}
+More information needed
 
-nlp(
-    "https://miro.medium.com/max/787/1*iECQRIiOGTmEFLdWkVIH2g.jpeg",
-    "What is the purchase amount?"
-)
-# {'score': 0.9912159, 'answer': '$1,000,000,000', 'start': 97, 'end': 97}
+## Training procedure
 
-nlp(
-    "https://www.accountingcoach.com/wp-content/uploads/2013/10/income-statement-example@2x.png",
-    "What are the 2020 net sales?"
-)
-# {'score': 0.59147286, 'answer': '$ 3,750', 'start': 19, 'end': 20}
-```
+### Training hyperparameters
 
-**NOTE**: This model relies on a [model definition](https://github.com/huggingface/transformers/pull/18407) and [pipeline](https://github.com/huggingface/transformers/pull/18414) that are currently in review to be included in the transformers project. In the meantime, you'll have to use the `trust_remote_code=True` flag to run this model.
+The following hyperparameters were used during training:
+- optimizer: None
+- training_precision: float32
 
-## About us
+### Training results
 
-This model was created by the team at [Impira](https://www.impira.com/).
+
+
+### Framework versions
+
+- Transformers 4.22.0.dev0
+- TensorFlow 2.9.2
+- Datasets 2.4.0
+- Tokenizers 0.12.1
diff --git a/config.json b/config.json
index fdf1c52..239d106 100644
--- a/config.json
+++ b/config.json
@@ -1,15 +1,17 @@
 {
-  "attention_probs_dropout_prob": 0.1,
+  "_name_or_path": "impira/layoutlm-document-qa",
   "architectures": [
     "LayoutLMForQuestionAnswering"
   ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": 0,
+  "classifier_dropout": null,
   "custom_pipelines": {
     "document-question-answering": {
       "impl": "pipeline_document_question_answering.DocumentQuestionAnsweringPipeline",
       "pt": "AutoModelForQuestionAnswering"
     }
   },
-  "bos_token_id": 0,
   "eos_token_id": 2,
   "gradient_checkpointing": false,
   "hidden_act": "gelu",
@@ -26,7 +28,7 @@
   "pad_token_id": 1,
   "position_embedding_type": "absolute",
   "tokenizer_class": "RobertaTokenizer",
-  "transformers_version": "4.6.1",
+  "transformers_version": "4.22.0.dev0",
   "type_vocab_size": 1,
   "use_cache": true,
   "vocab_size": 50265
diff --git a/tf_model.h5 b/tf_model.h5
new file mode 100644
index 0000000..b253626
--- /dev/null
+++ b/tf_model.h5
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2796bc2f67ac8e1abe55decfa104c7182376c4bf1f8b97ab87fd8bb4768f2f07
+size 511465184