From f4db109882d3bba8caa0313306545f944ae747b8 Mon Sep 17 00:00:00 2001
From: Lev <agsfer@gmail.com>
Date: Mon, 8 Jul 2024 17:03:20 +0300
Subject: [PATCH] fixies (#1313)

---
 docs/_data/navigation.yml                     |  2 +-
 docs/_includes/docs-sparckocr-pagination.html | 11 +++++
 docs/_includes/scripts/article.js             |  2 -
 docs/en/legal_release_notes.md                |  2 +-
 docs/en/licensed_version_compatibility.md     |  4 +-
 docs/en/ocr_benchmark.md                      |  2 +-
 docs/en/ocr_structures.md                     |  1 +
 .../release_notes_2_7_1.md                    |  1 +
 .../release_notes_4_4_0.md                    |  6 +--
 .../spark_ocr_versions/ocr_release_notes.md   |  6 +--
 .../release_notes_1_10_0.md                   |  4 ++
 .../release_notes_1_11_0.md                   |  8 ++++
 .../spark_ocr_versions/release_notes_1_2_0.md |  4 ++
 .../spark_ocr_versions/release_notes_1_3_0.md |  4 ++
 .../spark_ocr_versions/release_notes_1_4_0.md |  8 +++-
 .../spark_ocr_versions/release_notes_1_6_0.md |  1 +
 .../spark_ocr_versions/release_notes_1_8_0.md |  3 ++
 .../spark_ocr_versions/release_notes_1_9_0.md |  2 +
 .../spark_ocr_versions/release_notes_3_0_0.md |  4 ++
 .../release_notes_3_10_0.md                   |  3 ++
 .../release_notes_3_11_0.md                   |  8 +++-
 .../release_notes_3_12_0.md                   |  8 ++++
 .../release_notes_3_13_0.md                   |  7 ++-
 .../release_notes_3_14_0.md                   |  3 ++
 .../spark_ocr_versions/release_notes_3_1_0.md |  7 +++
 .../spark_ocr_versions/release_notes_3_2_0.md |  5 ++
 .../spark_ocr_versions/release_notes_3_3_0.md |  4 ++
 .../spark_ocr_versions/release_notes_3_4_0.md |  3 ++
 .../spark_ocr_versions/release_notes_3_5_0.md |  4 ++
 .../spark_ocr_versions/release_notes_3_6_0.md |  3 ++
 .../spark_ocr_versions/release_notes_3_7_0.md | 11 +++--
 .../spark_ocr_versions/release_notes_3_8_0.md |  3 ++
 .../spark_ocr_versions/release_notes_3_9_0.md |  6 +++
 .../spark_ocr_versions/release_notes_3_9_1.md |  4 ++
 .../spark_ocr_versions/release_notes_4_0_0.md |  8 ++++
 .../spark_ocr_versions/release_notes_4_0_2.md |  6 ++-
 .../spark_ocr_versions/release_notes_4_1_0.md |  9 ++--
 .../spark_ocr_versions/release_notes_4_2_0.md |  5 +-
 .../spark_ocr_versions/release_notes_4_2_1.md | 11 ++++-
 .../spark_ocr_versions/release_notes_4_2_4.md |  8 +++-
 .../spark_ocr_versions/release_notes_4_3_0.md |  4 ++
 .../spark_ocr_versions/release_notes_4_3_1.md |  8 ++++
 .../spark_ocr_versions/release_notes_4_3_3.md | 17 +++++--
 .../spark_ocr_versions/release_notes_4_4_0.md |  6 +++
 .../spark_ocr_versions/release_notes_4_4_1.md | 11 +++++
 .../spark_ocr_versions/release_notes_4_4_2.md |  6 ++-
 .../spark_ocr_versions/release_notes_4_4_3.md |  2 +
 .../spark_ocr_versions/release_notes_4_4_4.md |  2 +
 .../spark_ocr_versions/release_notes_5_0_0.md |  7 +++
 .../spark_ocr_versions/release_notes_5_0_1.md |  2 +-
 .../spark_ocr_versions/release_notes_5_0_2.md |  6 +--
 .../spark_ocr_versions/release_notes_5_1_0.md |  6 +--
 .../spark_ocr_versions/release_notes_5_1_2.md | 46 ++++++++-----------
 .../spark_ocr_versions/release_notes_5_2_0.md | 35 +++++++++-----
 .../spark_ocr_versions/release_notes_5_3_0.md |  6 ++-
 .../spark_ocr_versions/release_notes_5_3_1.md | 26 ++++++++---
 docs/en/version_compatibility.md              |  2 +-
 57 files changed, 298 insertions(+), 95 deletions(-)

diff --git a/docs/_data/navigation.yml b/docs/_data/navigation.yml
index 56a4ba6aca..f825b01cee 100755
--- a/docs/_data/navigation.yml
+++ b/docs/_data/navigation.yml
@@ -229,7 +229,7 @@ sparknlp-healthcare:
         url: /licensed/api/python
       - title: Wiki
         url: /docs/en/wiki
-      - title: Speed Benchmarks
+      - title: Benchmarks
         url: /docs/en/benchmark      
       - title: Best Practices Using Pretrained Models Together
         url: /docs/en/best_practices_pretrained_models
diff --git a/docs/_includes/docs-sparckocr-pagination.html b/docs/_includes/docs-sparckocr-pagination.html
index ac05da6997..d013668fc8 100644
--- a/docs/_includes/docs-sparckocr-pagination.html
+++ b/docs/_includes/docs-sparckocr-pagination.html
@@ -1,3 +1,14 @@
+<ul class="pagination">
+  <li>
+      <a href="#">Version <strong id="previosver"></strong></a>
+  </li>
+  <li>
+      <strong>Version <strong id="currversion"></strong></strong>
+  </li>
+  <li>
+      <a href="#">Version <strong id="nextver"></strong></a>
+  </li>
+</ul>
 <ul class="pagination owl-carousel pagination_big">
    <li><a href="release_notes_5_3_1">5.3.1</a></li>
     <li><a href="release_notes_5_3_0">5.3.0</a></li>
diff --git a/docs/_includes/scripts/article.js b/docs/_includes/scripts/article.js
index b0d8821d0d..3e2672171f 100755
--- a/docs/_includes/scripts/article.js
+++ b/docs/_includes/scripts/article.js
@@ -30,8 +30,6 @@ $(document).ready(function () {
       $('.pagination_big').owlCarousel({
         margin:10,
         nav:true,
-        center: true,
-        loop: true,
         dots:false,
         responsive:{
             0:{
diff --git a/docs/en/legal_release_notes.md b/docs/en/legal_release_notes.md
index 3fd1e98673..7a5306d520 100644
--- a/docs/en/legal_release_notes.md
+++ b/docs/en/legal_release_notes.md
@@ -15,7 +15,7 @@ sidebar:
 
 ## Releases log
 
-
+{:.table-model-big}
 | 	                                                                                                            |                                                                                                                                                      |                                                                                                             |                                                                                                        |
 |--------------------------------------------------------------------------------------------------------------|------------------------------------------------------------------------------------------------------------------------------------------------------|-------------------------------------------------------------------------------------------------------------|--------------------------------------------------------------------------------------------------------|
 | [1.0.0](https://medium.com/spark-nlp/spark-nlp-for-legal-1-0-0-over-300-new-state-of-the-art-models-in-multiple-languages-f3bae55c32e1)                         | [1.1.0](https://medium.com/@muhendisbp/legal-nlp-1-1-0-for-spark-nlp-has-been-released-89de7f099bdc)                                                 | [1.2.0](https://medium.com/spark-nlp/legal-nlp-1-2-0-for-spark-nlp-has-been-released-%EF%B8%8F-8d060b3391ef) | [1.3.0](https://gaddesaishailesh.medium.com/spark-nlp-for-legal-1-3-0-over-100-new-state-of-the-art-models-%EF%B8%8F-b069207ce77f) |
diff --git a/docs/en/licensed_version_compatibility.md b/docs/en/licensed_version_compatibility.md
index 66ba0047cf..d239ea74e7 100644
--- a/docs/en/licensed_version_compatibility.md
+++ b/docs/en/licensed_version_compatibility.md
@@ -13,7 +13,7 @@ sidebar:
 
 <div class="h3-box" markdown="1">
 
-
+{:.table-model-big}
 | Spark NLP for Healthcare	| Spark NLP (Public) |
 |---------------------------|--------------------|
 | 5.3.3                     | 5.3.2              |
@@ -95,7 +95,7 @@ sidebar:
 | 2.3.4   			        | 2.3.4              |
 
 
-
+{:.table-model-big}
 | Spark NLP for Healthcare	| Spark OCR          |
 |---------------------------|--------------------|
 | 4.3.0                     | 4.3.1              |
diff --git a/docs/en/ocr_benchmark.md b/docs/en/ocr_benchmark.md
index 7e1f13334c..86e70e9a09 100644
--- a/docs/en/ocr_benchmark.md
+++ b/docs/en/ocr_benchmark.md
@@ -33,7 +33,7 @@ sidebar:
 
 #### Benchmark Table
 
-{:.table-model-big.db}
+{:.table-model-big}
 | Instance      | memory | cores | input\_data\_pages| partition     | second per page | timing  |
 | ------------- | ------ | ----- | ----------------- | ------------- | --------------- | ------- |
 | m5n.4xlarge   | 64 GB  | 16    | 1000              | 10            | 0.24            | 4 mins  |
diff --git a/docs/en/ocr_structures.md b/docs/en/ocr_structures.md
index a9abb6303d..319e42a200 100644
--- a/docs/en/ocr_structures.md
+++ b/docs/en/ocr_structures.md
@@ -367,6 +367,7 @@ Show single image with metadata in Jupyter notebook.
 
 {:.table-model-big}
 | Param name | Type | Default | Description |
+|------------|------|---------|-------------|
 | width | string | "600" | width of image |
 | show_meta | boolean | true | enable/disable displaying metadata of image |
 
diff --git a/docs/en/spark_nlp_healthcare_versions/release_notes_2_7_1.md b/docs/en/spark_nlp_healthcare_versions/release_notes_2_7_1.md
index 7993c99bc7..c8a99abad9 100644
--- a/docs/en/spark_nlp_healthcare_versions/release_notes_2_7_1.md
+++ b/docs/en/spark_nlp_healthcare_versions/release_notes_2_7_1.md
@@ -79,6 +79,7 @@ Output:
 
 {:.table-model-big}
 |    | chunks                      |   begin |   end |      code | resolutions
+|----|-----------------------------|---------|-------|-----------|------------|
 |  2 | COPD  				 |     113 |   116 |  13645005 | copd - chronic obstructive pulmonary disease
 |  8 | PTCA                        |     324 |   327 | 373108000 | post percutaneous transluminal coronary angioplasty (finding)
 | 16 | close monitoring            |     519 |   534 | 417014005 | on examination - vigilance
diff --git a/docs/en/spark_nlp_healthcare_versions/release_notes_4_4_0.md b/docs/en/spark_nlp_healthcare_versions/release_notes_4_4_0.md
index a25b3a1f74..8cd84d3037 100644
--- a/docs/en/spark_nlp_healthcare_versions/release_notes_4_4_0.md
+++ b/docs/en/spark_nlp_healthcare_versions/release_notes_4_4_0.md
@@ -88,7 +88,7 @@ Our clinical summarizer models with only 250M parameters perform 30-35% better t
 
 🔎 Benchmark on MtSamples Summarization Dataset
 
-{:.table-model-big}
+{:.table-model-big.db}
 | model_name | model_size | Rouge | Bleu | bertscore_precision | bertscore_recall: | bertscore_f1 |
 |--|--|--|--|--|--|--|
 philschmid/flan-t5-base-samsum | 250M | 0.1919 | 0.1124 | 0.8409 | 0.8964 | 0.8678 |
@@ -100,7 +100,7 @@ transformersbook/pegasus-samsum | 500M | 0.1924 | 0.0965 | 0.8920 | 0.8149 | 0.8
 
 🔎 Benchmark on MIMIC Summarization Dataset
 
-{:.table-model-big}
+{:.table-model-big.db}
 | model_name | model_size | Rouge | Bleu | bertscore_precision | bertscore_recall: | bertscore_f1 |
 |--|--|--|--|--|--|--|
 philschmid/flan-t5-base-samsum | 250M | 0.1910 | 0.1037 | 0.8708 | 0.9056 | 0.8879 |
@@ -110,7 +110,7 @@ transformersbook/pegasus-samsum | 570M | 0.1425 | 0.0582 | 0.9171 | 0.8682 | 0.8
 **summarizer_clinical_jsl** | **250M** | **0.395** | **0.2962** | **0.895** | **0.9316** | **0.913** |
 **summarizer_clinical_jsl_augmented** | **250M** | **0.3964** | **0.307** | **0.9109** | **0.9452** | **0.9227** |
 
-![image](https://user-images.githubusercontent.com/64752006/230899745-3a67d142-1bdf-4f4b-83cb-d012953b1e09.png)
+![Benchmark on MIMIC Summarization Dataset](https://user-images.githubusercontent.com/64752006/230899745-3a67d142-1bdf-4f4b-83cb-d012953b1e09.png)
 
 </div><div class="h3-box" markdown="1">
 
diff --git a/docs/en/spark_ocr_versions/ocr_release_notes.md b/docs/en/spark_ocr_versions/ocr_release_notes.md
index 28694f5d28..c43100632a 100644
--- a/docs/en/spark_ocr_versions/ocr_release_notes.md
+++ b/docs/en/spark_ocr_versions/ocr_release_notes.md
@@ -39,6 +39,7 @@ Release date: 11-04-2024
 ## Improved table extraction capabilities in HocrToTextTable
 Many issues related to column detection in our Table Extraction pipelines are addressed in this release, compared to previous Visual NLP version the metrics have improved. Table below shows F1-score(CAR or Cell Adjacency Relationship) performances on ICDAR 19 Track B dataset for different IoU values of our two versions in comparison with [other results](https://paperswithcode.com/paper/multi-type-td-tsr-extracting-tables-from/review/).
 
+{:.table-model-big}
 | Model  | 0.6 | 0.7 | 0.8 | 0.9 |
 | ------------- | ------------- |------------- |------------- |------------- |
 | CascadeTabNet	  | 0.438  | 0.354 | 0.19 | 0.036 |
@@ -88,7 +89,7 @@ ocr = ImageToTextV2.pretrained("ocr_base_printed_v2_opt", "en", "clinical/ocr")
     .setIncludeConfidence(True)
 ```
 
-![image](/assets/images/ocr/confidence_score.png)
+![Confidence scores in ImageToTextV2](/assets/images/ocr/confidence_score.png)
 
 Check this [updated notebook](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/master/jupyter/TextRecognition/SparkOcrImageToTextV2.ipynb) for an end-to-end example.
 
@@ -131,9 +132,6 @@ ImageDrawRegions is the annotator used for rendering regions into images so we c
 ### Bug Fixes
 + PdfToImage resetting page information when used in the same pipeline as PdfToText: When the sequence {PdfToText, PdfToImage} was used the original pages computed at PdfToText where resetted to zero by PdfToImage.
 
-
-
-
 </div><div class="prev_ver h3-box" markdown="1">
 
 ## Previous versions
diff --git a/docs/en/spark_ocr_versions/release_notes_1_10_0.md b/docs/en/spark_ocr_versions/release_notes_1_10_0.md
index 411af13b81..ed2fcb125f 100644
--- a/docs/en/spark_ocr_versions/release_notes_1_10_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_1_10_0.md
@@ -21,6 +21,8 @@ Release date: 20-01-2021
 
 Support Microsoft Docx documents.
 
+</div><div class="h3-box" markdown="1">
+
 #### New Features
 
 * Added [DocToText](/docs/en/ocr_pipeline_components#doctotext) transformer for extract text
@@ -30,6 +32,8 @@ table data from DOCX documents.
 * Added [DocToPdf](/docs/en/ocr_pipeline_components#doctopdf) transformer for convert DOCX
  documents to PDF format.
 
+</div><div class="h3-box" markdown="1">
+
 #### Bugfixes
 
 * Fixed issue with loading model data on some cluster configurations
diff --git a/docs/en/spark_ocr_versions/release_notes_1_11_0.md b/docs/en/spark_ocr_versions/release_notes_1_11_0.md
index 6663a4d190..ed144eec00 100644
--- a/docs/en/spark_ocr_versions/release_notes_1_11_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_1_11_0.md
@@ -22,22 +22,30 @@ Release date: 25-02-2021
 Support German, French, Spanish and Russian languages.
 Improving [PositionsFinder](/docs/en/ocr_pipeline_components#positionsfinder) and ImageToText for better support de-identification.
 
+</div><div class="h3-box" markdown="1">
+
 #### New Features
 
 * Loading model data from S3 in [ImageToText](/docs/en/ocr_pipeline_components#imagetotext).
 * Added support German, French, Spanish, Russian languages in [ImageToText](/docs/en/ocr_pipeline_components#imagetotext).
 * Added different OCR model types: Base, Best, Fast in [ImageToText](/docs/en/ocr_pipeline_components#imagetotext).
 
+</div><div class="h3-box" markdown="1">
+
 #### Enhancements
 
 * Added spaces symbols to the output positions in the [ImageToText](/docs/en/ocr_pipeline_components#imagetotext) transformer.
 * Eliminate python-levensthein from dependencies for simplify installation.
 
+</div><div class="h3-box" markdown="1">
+
 #### Bugfixes
 
 * Fixed issue with extracting coordinates in  in [ImageToText](/docs/en/ocr_pipeline_components#imagetotext).
 * Fixed loading model data on cluster in yarn mode.
 
+</div><div class="h3-box" markdown="1">
+
 #### New notebooks
 
 * [Languages Support](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/1.11.0/jupyter/SparkOcrLanguagesSupport.ipynb)
diff --git a/docs/en/spark_ocr_versions/release_notes_1_2_0.md b/docs/en/spark_ocr_versions/release_notes_1_2_0.md
index 8a2a96cb7b..160ef4dc21 100644
--- a/docs/en/spark_ocr_versions/release_notes_1_2_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_1_2_0.md
@@ -22,6 +22,8 @@ Release date: 08-04-2020
 
 Improved support Databricks and processing selectable pdfs.
 
+</div><div class="h3-box" markdown="1">
+
 #### Enhancements
 
 * Adapted Spark OCR for run on Databricks.
@@ -29,6 +31,8 @@ Improved support Databricks and processing selectable pdfs.
 * Added 'positionsCol' param to [ImageToText](/docs/en/ocr_pipeline_components#imagetotext).
 * Improved support Spark NLP. Changed [start](/docs/en/ocr_install#using-start-function) function.
 
+</div><div class="h3-box" markdown="1">
+
 #### New Features
 
 * Added [showImage](/docs/en/ocr_structures#showimages) implicit to Dataframe for display images in Scala Databricks notebooks.
diff --git a/docs/en/spark_ocr_versions/release_notes_1_3_0.md b/docs/en/spark_ocr_versions/release_notes_1_3_0.md
index 2c04811d14..f77c7e7628 100644
--- a/docs/en/spark_ocr_versions/release_notes_1_3_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_1_3_0.md
@@ -21,12 +21,16 @@ Release date: 22-05-2020
 
 New functionality for de-identification problem.
 
+</div><div class="h3-box" markdown="1">
+
 #### Enhancements
 
 * Renamed TesseractOCR to ImageToText. 
 * Simplified installation.
 * Added check license from `SPARK_NLP_LICENSE` env varibale.
 
+</div><div class="h3-box" markdown="1">
+
 #### New Features
 
 * Support storing for binaryFormat. Added support storing Image and PDF files.
diff --git a/docs/en/spark_ocr_versions/release_notes_1_4_0.md b/docs/en/spark_ocr_versions/release_notes_1_4_0.md
index ea9689dea9..184cedf5d2 100644
--- a/docs/en/spark_ocr_versions/release_notes_1_4_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_1_4_0.md
@@ -21,6 +21,8 @@ Release date: 23-06-2020
 
 Added support Dicom format and improved support image morphological operations.
 
+</div><div class="h3-box" markdown="1">
+
 #### Enhancements
 
 * Updated [start](/docs/en/ocr_install#using-start-function) function. Improved support Spark NLP internal.
@@ -28,12 +30,16 @@ Added support Dicom format and improved support image morphological operations.
 * Improved existing transformers for support de-identification Dicom documents.
 * Added possibility to draw filled rectangles to [ImageDrawRegions](/docs/en/ocr_pipeline_components#imagedrawregions).
 
+</div><div class="h3-box" markdown="1">
+
 #### New Features
 
 * Support reading and writing Dicom documents.
 * Added [ImageMorphologyOperation](/docs/en/ocr_pipeline_components#imagemorphologyoperation) transformer which support:
  erosion, dilation, opening and closing operations.
- 
+
+</div><div class="h3-box" markdown="1">
+
 #### Bugfixes
 
 * Fixed issue in [ImageToText](/docs/en/ocr_pipeline_components#imagetotext) related to extraction coordinates.
diff --git a/docs/en/spark_ocr_versions/release_notes_1_6_0.md b/docs/en/spark_ocr_versions/release_notes_1_6_0.md
index ab215b3d72..b1277e4ea1 100644
--- a/docs/en/spark_ocr_versions/release_notes_1_6_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_1_6_0.md
@@ -21,6 +21,7 @@ Release date: 05-09-2020
 
 Support parsing data from tables for selectable PDFs.
 
+</div><div class="h3-box" markdown="1">
 
 #### New Features
 
diff --git a/docs/en/spark_ocr_versions/release_notes_1_8_0.md b/docs/en/spark_ocr_versions/release_notes_1_8_0.md
index 12c5d92499..d2a231e5ce 100644
--- a/docs/en/spark_ocr_versions/release_notes_1_8_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_1_8_0.md
@@ -22,6 +22,8 @@ Release date: 20-11-2020
 Optimisation performance for processing multipage PDF documents.
 Support up to 10k pages per document.
 
+</div><div class="h3-box" markdown="1">
+
 #### New Features
 
 * Added [ImageAdaptiveBinarizer](/docs/en/ocr_pipeline_components#imageadaptivebinarizer) Scala transformer with support:
@@ -30,6 +32,7 @@ Support up to 10k pages per document.
     - Sauvola local thresholding
 * Added possibility to split pdf to small documents for optimize processing in [PdfToImage](/docs/en/ocr_pipeline_components#pdftoimage).
 
+</div><div class="h3-box" markdown="1">
 
 #### Enhancements
 
diff --git a/docs/en/spark_ocr_versions/release_notes_1_9_0.md b/docs/en/spark_ocr_versions/release_notes_1_9_0.md
index 49412af70b..2ddfc6404f 100644
--- a/docs/en/spark_ocr_versions/release_notes_1_9_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_1_9_0.md
@@ -21,6 +21,8 @@ Release date: 11-12-2020
 
 Extension of  FoundationOne report parser and support HOCR output format.
 
+</div><div class="h3-box" markdown="1">
+
 #### New Features
 
 * Added [ImageToHocr](/docs/en/ocr_pipeline_components#imagetohocr) transformer for recognize text from image and store it to HOCR format.
diff --git a/docs/en/spark_ocr_versions/release_notes_3_0_0.md b/docs/en/spark_ocr_versions/release_notes_3_0_0.md
index e14872b00a..9713fed3bd 100644
--- a/docs/en/spark_ocr_versions/release_notes_3_0_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_3_0_0.md
@@ -25,6 +25,8 @@ Spark OCR 3.0.0 extends the support for Apache Spark 3.0.x and 3.1.x major relea
 
 Spark OCR started to support Tensorflow models. First model is [VisualDocumentClassifier](/docs/en/ocr_pipeline_components#visualdocumentclassifier).
 
+</div><div class="h3-box" markdown="1">
+
 #### New Features
 
 * Support for Apache Spark and PySpark 3.0.x on Scala 2.12
@@ -47,6 +49,8 @@ Spark OCR started to support Tensorflow models. First model is [VisualDocumentCl
 * [VisualDocumentClassifier](/docs/en/ocr_pipeline_components#visualdocumentclassifier) model for classification documents using text and layout data.
 * Added support Vietnamese language.
 
+</div><div class="h3-box" markdown="1">
+
 #### New notebooks
 
 * [Visual Document Classifier](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/master/jupyter/SparkOCRVisualDocumentClassifier.ipynb)
diff --git a/docs/en/spark_ocr_versions/release_notes_3_10_0.md b/docs/en/spark_ocr_versions/release_notes_3_10_0.md
index 83f534d77c..5267d12e4e 100644
--- a/docs/en/spark_ocr_versions/release_notes_3_10_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_3_10_0.md
@@ -22,6 +22,7 @@ Release date: 10-01-2022
 
 Form recognition using LayoutLMv2 and text detection.
 
+</div><div class="h3-box" markdown="1">
 
 #### New Features
 
@@ -30,12 +31,14 @@ Form recognition using LayoutLMv2 and text detection.
 * Support rotated regions in [ImageSplitRegions](/docs/en/ocr_pipeline_components#imagesplitregions)
 * Support rotated regions in [ImageDrawRegions](/docs/en/ocr_pipeline_components#imagedrawregions)
 
+</div><div class="h3-box" markdown="1">
 
 #### New Models
 
 * LayoutLMv2 fine-tuned on FUNSD dataset
 * Text detection model based on CRAFT architecture
 
+</div><div class="h3-box" markdown="1">
 
 #### New notebooks
 
diff --git a/docs/en/spark_ocr_versions/release_notes_3_11_0.md b/docs/en/spark_ocr_versions/release_notes_3_11_0.md
index 98c2a907f5..af98ec7839 100644
--- a/docs/en/spark_ocr_versions/release_notes_3_11_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_3_11_0.md
@@ -20,9 +20,11 @@ Release date: 28-02-2022
 
 #### Overview
 
-We are glad to announce that Spark OCR 3.11.0 has been released!.
+We are glad to announce that Spark OCR 3.11.0 has been released!
 This release comes with new models, new features, bug fixes, and notebook examples.
 
+</div><div class="h3-box" markdown="1">
+
 #### New Features
 
 * Added [ImageTextDetectorV2](/docs/en/ocr_object_detection#imagetextdetectorv2) Python Spark-OCR Transformer for detecting printed and handwritten text
@@ -32,11 +34,15 @@ This release comes with new models, new features, bug fixes, and notebook exampl
 * Added [FormRelationExtractor](/docs/en/ocr_visual_document_understanding#formrelationextractor) for detecting relations between key and value entities in forms.
 * Added the capability of fine tuning VisualDocumentNerV2 models for key-value pairs extraction.
 
+</div><div class="h3-box" markdown="1">
+
 #### New Models
 
 * ImageTextDetectorV2: this extends the ImageTextDetectorV1 character level text detection model with a refiner net architecture.
 * ImageTextRecognizerV2: Text recognition for printed text based on the Deep Learning Transformer Architecture.
 
+</div><div class="h3-box" markdown="1">
+
 #### New notebooks
 
 * [SparkOcrImageToTextV2](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/3110-release-candidate/jupyter/TextRecognition/SparkOcrImageToTextV2.ipynb)
diff --git a/docs/en/spark_ocr_versions/release_notes_3_12_0.md b/docs/en/spark_ocr_versions/release_notes_3_12_0.md
index af1b1ba3cb..d68d92be58 100644
--- a/docs/en/spark_ocr_versions/release_notes_3_12_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_3_12_0.md
@@ -21,6 +21,8 @@ Release date: 14-04-2022
 We're glad to announce that Spark OCR 3.12.0 has been released!
 This release comes with new models for Handwritten Text Recognition, Spark 3.2 support, bug fixes, and notebook examples.
 
+</div><div class="h3-box" markdown="1">
+
 #### New Features
 
 * Added to the ImageTextDetectorV2:
@@ -57,16 +59,22 @@ spark.conf.set("spark.sql.optimizer.nestedSchemaPruning.enabled", False)
 
 * Improved documentation on the website.
 
+</div><div class="h3-box" markdown="1">
+
 #### New Models
 
 ocr_small_printed: Text recognition small model for printed text based on ImageToTextV2
 ocr_small_handwritten: Text recognition small model for handwritten text based on ImageToTextV2
 ocr_base_handwritten: Text recognition base model for handwritten text based on ImageToTextV2
 
+</div><div class="h3-box" markdown="1">
+
 #### Bug Fixes
 
 * display_table() function failing to display tables coming from digital PDFs.
 
+</div><div class="h3-box" markdown="1">
+
 #### New notebooks
 
 * [SparkOcrImageToTextV2OutputFormats.ipynb](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/3120-release-candidate/jupyter/TextRecognition/SparkOcrImageToTextV2OutputFormats.ipynb), different output formats for ImageToTextV2.
diff --git a/docs/en/spark_ocr_versions/release_notes_3_13_0.md b/docs/en/spark_ocr_versions/release_notes_3_13_0.md
index a2ae6212a0..25f6f04621 100644
--- a/docs/en/spark_ocr_versions/release_notes_3_13_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_3_13_0.md
@@ -17,9 +17,11 @@ sidebar:
 
 Release date: 25-05-2022
 
-We are glad to announce that Spark OCR 3.13.0 has been released!.
+We are glad to announce that Spark OCR 3.13.0 has been released!
 This release focuses around VisualDocumentNer models, adding ability to fine-tune, fixing bugs, and to leverage the Annotation Lab to generate training data.
 
+</div><div class="h3-box" markdown="1">
+
 #### New Features
 
 * VisualDocumentNerV21:
@@ -28,11 +30,14 @@ This release focuses around VisualDocumentNer models, adding ability to fine-tun
 * AlabReaders: 
   * New class to allow training data from the Annotation Lab to be imported into Spark OCR. Currently, the reader supports Visual Ner only.
 
+</div><div class="h3-box" markdown="1">
 
 #### Bug Fixes
 
 * Feature extraction on VisualDocumentNer has been improved.
 
+</div><div class="h3-box" markdown="1">
+
 #### New notebooks
 
 * [SparkOcrFormRecognitionFineTuning.ipynb](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/3.13.0-release-candidate/jupyter/FormRecognition/SparkOcrFormRecognitionFineTuning.ipynb), end to end example on Visual Document Ner Fine-Tuning.
diff --git a/docs/en/spark_ocr_versions/release_notes_3_14_0.md b/docs/en/spark_ocr_versions/release_notes_3_14_0.md
index 5afd20f018..8ba8ca8e3a 100644
--- a/docs/en/spark_ocr_versions/release_notes_3_14_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_3_14_0.md
@@ -22,6 +22,8 @@ Release date: 13-06-2022
 We are glad to announce that Spark OCR 3.14.0 has been released!.
 This release focuses around Visual Document Classification models, native Image Preprocessing on the JVM, and bug fixes.
 
+</div><div class="h3-box" markdown="1">
+
 #### New Features
 
 * VisualDocumentClassifierv2:
@@ -33,6 +35,7 @@ This release focuses around Visual Document Classification models, native Image
 * ImageTransformer:
   * New transformer that provides different image transformations on the JVM. Supported transforms are Scaling, Adaptive Thresholding, Median Blur, Dilation, Erosion, and Object Removal.
 
+</div><div class="h3-box" markdown="1">
 
 #### New notebooks
 
diff --git a/docs/en/spark_ocr_versions/release_notes_3_1_0.md b/docs/en/spark_ocr_versions/release_notes_3_1_0.md
index 9c4c362d6d..ddcfc6345e 100644
--- a/docs/en/spark_ocr_versions/release_notes_3_1_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_3_1_0.md
@@ -23,20 +23,27 @@ Image processing on GPU. It is in 3.5 times faster than on CPU.
 
 More details please read in [GPU image preprocessing in Spark OCR](https://medium.com/spark-nlp/gpu-image-pre-processing-in-spark-ocr-3-1-0-6fc27560a9bb)
 
+</div><div class="h3-box" markdown="1">
 
 #### New Features
 
 * [GPUImageTransformer](/docs/en/ocr_pipeline_components#gpuimagetransformer) with support: scaling, erosion, delation, Otsu and Huang thresholding.
 * Added [display_images](/docs/en/ocr_structures#displayimages) util function for displaying images from Spark DataFrame in Jupyter notebooks.
 
+</div><div class="h3-box" markdown="1">
+
 #### Enhancements
 
 * Improve [display_image](/docs/en/ocr_structures#displayimage) util function.
 
+</div><div class="h3-box" markdown="1">
+
 #### Bug fixes
 
 * Fixed issue with extra dependencies in [start](/docs/en/ocr_install#using-start-function) function
 
+</div><div class="h3-box" markdown="1">
+
 #### New notebooks
 
 * [GPU image processing](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/3.1.0/jupyter/SparkOCRGPUOperations.ipynb)
diff --git a/docs/en/spark_ocr_versions/release_notes_3_2_0.md b/docs/en/spark_ocr_versions/release_notes_3_2_0.md
index a173f87eef..85a1aba2c6 100644
--- a/docs/en/spark_ocr_versions/release_notes_3_2_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_3_2_0.md
@@ -23,23 +23,28 @@ Multi-modal visual document understanding, built on the LayoutLM architecture.
 It achieves new state-of-the-art accuracy in several downstream tasks,
 including form understanding and receipt understanding.
 
+</div><div class="h3-box" markdown="1">
 
 #### New Features
 
 * [VisualDocumentNER](/docs/en/ocr_pipeline_components#visualdocumentner) is a DL model for NER problem using text and layout data.
   Currently available pre-trained model on the SROIE dataset.
 
+</div><div class="h3-box" markdown="1">
 
 #### Enhancements
 
 * Added support `SPARK_OCR_LICENSE` env key for read license.
 * Update dependencies and sync Spark versions with Spark NLP.
 
+</div><div class="h3-box" markdown="1">
 
 #### Bugfixes
 
 * Fixed an issue that some ImageReaderSpi plugins are unavailable in the fat jar.
 
+</div><div class="h3-box" markdown="1">
+
 #### New notebooks
 
 * [Visual Document NER](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/3.2.0/jupyter/SparkOCRVisualDocumentNer.ipynb)
diff --git a/docs/en/spark_ocr_versions/release_notes_3_3_0.md b/docs/en/spark_ocr_versions/release_notes_3_3_0.md
index e1ac4de870..d01422d5cb 100644
--- a/docs/en/spark_ocr_versions/release_notes_3_3_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_3_3_0.md
@@ -27,12 +27,16 @@ The model was pre-trained on the __COCO dataset__ and fine-tuned on __ICDAR 2019
 
 More details please read in [Table Detection & Extraction in Spark OCR](https://medium.com/spark-nlp/table-detection-extraction-in-spark-ocr-50765c6cedc9)
 
+</div><div class="h3-box" markdown="1">
+
 #### New Features
 
 * [ImageTableDetector](/docs/en/ocr_table_recognition#imagetabledetector) is a DL model for detect tables on the image.
 * [ImageTableCellDetector](/docs/en/ocr_table_recognition#imagetablecelldetector) is a transformer for detect regions of cells in the table image.
 * [ImageCellsToTextTable](/docs/en/ocr_table_recognition#imagecellstotexttable) is a transformer for extract text from the detected cells.
 
+</div><div class="h3-box" markdown="1">
+
 #### New notebooks
 
 * [Image Table Detection example](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/3.3.0/jupyter/SparkOcrImageTableDetection.ipynb)
diff --git a/docs/en/spark_ocr_versions/release_notes_3_4_0.md b/docs/en/spark_ocr_versions/release_notes_3_4_0.md
index a1ebffa836..df8542e427 100644
--- a/docs/en/spark_ocr_versions/release_notes_3_4_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_3_4_0.md
@@ -23,10 +23,13 @@ Signature Detection in image-based documents.
 
 More details please read in [Signature Detection in Spark OCR](https://medium.com/spark-nlp/signature-detection-in-spark-ocr-32f9e6f91e3c)
 
+</div><div class="h3-box" markdown="1">
+
 #### New Features
 
 * [ImageSignatureDetector](/docs/en/ocr_object_detection#imagehandwrittendetector) is a DL model for detecting signature on the image.
 
+</div><div class="h3-box" markdown="1">
 
 #### New notebooks
 
diff --git a/docs/en/spark_ocr_versions/release_notes_3_5_0.md b/docs/en/spark_ocr_versions/release_notes_3_5_0.md
index a26b0177b9..55d60be280 100644
--- a/docs/en/spark_ocr_versions/release_notes_3_5_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_3_5_0.md
@@ -23,6 +23,7 @@ Improve table detection and table recognition.
 
 More details please read in [Extract Tabular Data from PDF in Spark OCR](https://medium.com/spark-nlp/extract-tabular-data-from-pdf-in-spark-ocr-b02136bc0fcb)
 
+</div><div class="h3-box" markdown="1">
 
 #### New Features
 
@@ -30,12 +31,15 @@ More details please read in [Extract Tabular Data from PDF in Spark OCR](https:/
 borderless tables and combined tables.
 * Added __Wolf__ and __Singh__ adaptive binarization methods to the [ImageAdaptiveThresholding](../ocr_pipeline_components#imageadaptivethresholding).
 
+</div><div class="h3-box" markdown="1">
 
 #### Enhancements
 
 * Added possibility to use different type of images as input for [ImageTableDetector](/docs/en/ocr_table_recognition#imagetabledetector).
 * Added [display_pdf](/docs/en/ocr_structures#displaypdf) and [display_images_horizontal](/docs/en/ocr_structures#displayimageshorizontal) util functions.
 
+</div><div class="h3-box" markdown="1">
+
 #### New notebooks
 
 * [Tables Recognition from PDF](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/3.5.0/jupyter/SparkOcrImageTableRecognitionPdf.ipynb)
diff --git a/docs/en/spark_ocr_versions/release_notes_3_6_0.md b/docs/en/spark_ocr_versions/release_notes_3_6_0.md
index 07a5e581d3..b5c58ceb24 100644
--- a/docs/en/spark_ocr_versions/release_notes_3_6_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_3_6_0.md
@@ -30,6 +30,7 @@ Handwritten detection and visualization improvement.
 * Added possibility to scale image to fixed size in [ImageScaler](/docs/en/ocr_pipeline_components#imagescaler)
  with keeping original ratio.
 
+</div><div class="h3-box" markdown="1">
 
 #### Enhancements
 
@@ -38,11 +39,13 @@ Handwritten detection and visualization improvement.
 * Added an auto inferencing of the input shape for detector models
 * New license validator
 
+</div><div class="h3-box" markdown="1">
 
 #### Bugfixes
 
 * Fixed display BGR images in display functions
 
+</div><div class="h3-box" markdown="1">
 
 #### New and updated notebooks
 
diff --git a/docs/en/spark_ocr_versions/release_notes_3_7_0.md b/docs/en/spark_ocr_versions/release_notes_3_7_0.md
index 8b40c001c0..44fbbc3fe7 100644
--- a/docs/en/spark_ocr_versions/release_notes_3_7_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_3_7_0.md
@@ -19,16 +19,16 @@ Release date: 30-08-2021
 
 #### Overview
 
-Improve table recognition and render OCR results to the PDF with original image
+Improve table recognition and render OCR results to the PDF with original image.
 
+</div><div class="h3-box" markdown="1">
 
 #### New Features
 
-* Added [ImageToTextPdf](/docs/en/ocr_pipeline_components#imagetotextpdf) transformer for storing recognized text to the searchable
-PDF with original image
-* Added [PdfAssembler](/docs/en/ocr_pipeline_components#pdfassembler) for assembling multipage PDF document from single page PDF
-documents
+* Added [ImageToTextPdf](/docs/en/ocr_pipeline_components#imagetotextpdf) transformer for storing recognized text to the searchable PDF with original image
+* Added [PdfAssembler](/docs/en/ocr_pipeline_components#pdfassembler) for assembling multipage PDF document from single page PDF documents
 
+</div><div class="h3-box" markdown="1">
 
 #### Enhancements
 
@@ -41,6 +41,7 @@ documents
 * Added [display_pdf_file](/docs/en/ocr_structures#displaypdffile) function for displaying pdf in embedded pdf viewer
 * Updated license validator
 
+</div><div class="h3-box" markdown="1">
 
 #### New and updated notebooks
 
diff --git a/docs/en/spark_ocr_versions/release_notes_3_8_0.md b/docs/en/spark_ocr_versions/release_notes_3_8_0.md
index 2f5fb36ecd..5fd63b23a8 100644
--- a/docs/en/spark_ocr_versions/release_notes_3_8_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_3_8_0.md
@@ -21,11 +21,14 @@ Release date: 14-09-2021
 
 Support Microsoft PPT and PPTX documents.
 
+</div><div class="h3-box" markdown="1">
+
 #### New Features
 
 * Added [PptToPdf](/docs/en/ocr_pipeline_components#ppttopdf) transformer for converting PPT and PPTX slides to the PDF document.
 * Added [PptToTextTable](/docs/en/ocr_pipeline_components#ppttotexttable) transformer for extracting tables from PPT and PPTX slides.
 
+</div><div class="h3-box" markdown="1">
 
 #### New and updated notebooks
 
diff --git a/docs/en/spark_ocr_versions/release_notes_3_9_0.md b/docs/en/spark_ocr_versions/release_notes_3_9_0.md
index 802e9bed88..ebc8da184e 100644
--- a/docs/en/spark_ocr_versions/release_notes_3_9_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_3_9_0.md
@@ -21,6 +21,8 @@ Release date: 20-10-2021
 
 Improve visualization and support Spark NLP. 
 
+</div><div class="h3-box" markdown="1">
+
 #### New Features
 
 * Added [HocrTokenizer](/docs/en/ocr_pipeline_components#hocrtokenizer)
@@ -28,6 +30,8 @@ Improve visualization and support Spark NLP.
 * Added [ImageDrawAnnotations](/docs/en/ocr_pipeline_components#imagedrawannotations)
 * Added support Arabic language in ImageToText and ImageToHocr
 
+</div><div class="h3-box" markdown="1">
+
 #### Enhancements
 
 * Added postprocessing to the [ImageTableDetector](/docs/en/ocr_table_recognition#imagetabledetector)
@@ -36,6 +40,8 @@ Improve visualization and support Spark NLP.
 * Updated license-validator. Added support floating license and set AWS keys from license.
 * Added 'whiteList' param to the [VisualDocumentNER](/docs/en/ocr_pipeline_components#visualdocumentner)
 
+</div><div class="h3-box" markdown="1">
+
 #### New and updated notebooks
 
 * [Spark OCR HOCR](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/3.9.0/jupyter/SparkOcrHocr.ipynb)
diff --git a/docs/en/spark_ocr_versions/release_notes_3_9_1.md b/docs/en/spark_ocr_versions/release_notes_3_9_1.md
index 133f38eb85..69855afe06 100644
--- a/docs/en/spark_ocr_versions/release_notes_3_9_1.md
+++ b/docs/en/spark_ocr_versions/release_notes_3_9_1.md
@@ -21,10 +21,14 @@ Release date: 02-11-2021
 
 Added preserving of original file formatting
 
+</div><div class="h3-box" markdown="1">
+
 #### Enhancements
 
 * Added keepLayout param to the [ImageToText](/docs/en/ocr_pipeline_components#imagetotext)
 
+</div><div class="h3-box" markdown="1">
+
 #### New and updated notebooks
 
 * [Preserve Original Formatting](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/3.9.1/jupyter/SparkOcrPreserveOriginalFormatting.ipynb)
diff --git a/docs/en/spark_ocr_versions/release_notes_4_0_0.md b/docs/en/spark_ocr_versions/release_notes_4_0_0.md
index 80ae004cd6..13815f3a38 100644
--- a/docs/en/spark_ocr_versions/release_notes_4_0_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_4_0_0.md
@@ -22,21 +22,29 @@ Release date: 16-07-2022
 We are very glad to announce that Spark OCR 4.0.0 has been released!
 This release comes with new models, new functionality, bug fixes, and compatibility with 4.0.0 versions of Spark NLP and Spark NLP for Healthcare.
 
+</div><div class="h3-box" markdown="1">
+
 #### New Features
 * New DicomMetadataDeidentifier class to help deidentifying metadata of dicom files. Example Notebook.
 * New helper function display_dicom() to help displaying DICOM files in notebooks.
 * New DicomDrawRegions that can clean burned pixels for removing PHI.
 * Improved support for DICOM files containing 12bit images.
 
+</div><div class="h3-box" markdown="1">
+
 #### Bug Fixes
 * Fixes on the Visual NER Finetuning process including VisualDocumentNERv2 and AlabReader.
 * Improved exception handling for VisualDocumentClassifier models.
 
+</div><div class="h3-box" markdown="1">
+
 #### New Models
 * New LayoutLMv3 based Visual Document NER: layoutlmv3_finetuned_funsd.
 * Improved handwritten detection ocr_base_handwritten_v2.
 * VisualDocumentClassifierV2: layoutlmv2_rvl_cdip_40k. This model adds more data compared to layoutlmv2_rvl_cdip_1500, and achieves an accuracy of 88%.
 
+</div><div class="h3-box" markdown="1">
+
 #### Compatibility Updates
 * Deprecated Spark 2.3 and Spark 2.4 support.
 * Tested compatibility with Spark-NLP and Spark NLP for Healthcare 4.0.0.
diff --git a/docs/en/spark_ocr_versions/release_notes_4_0_2.md b/docs/en/spark_ocr_versions/release_notes_4_0_2.md
index df55eaf2f3..95afcbb34e 100644
--- a/docs/en/spark_ocr_versions/release_notes_4_0_2.md
+++ b/docs/en/spark_ocr_versions/release_notes_4_0_2.md
@@ -20,8 +20,9 @@ Release date: 12-09-2022
 #### Overview
 
 We are glad to announce that Spark OCR 4.0.2 has been released!
-This release comes with new features, fixes and more!.
+This release comes with new features, fixes and more!
 
+</div><div class="h3-box" markdown="1">
 
 #### New Features
 
@@ -29,13 +30,14 @@ This release comes with new features, fixes and more!.
 * Added support for Simplified Chinese.
 * Added new 'PdfToForm' annotator, capable of extracting forms from digital PDFs. This is different from previously introduced VisualDocumentNER annotator in that this new annotator works only on digital documents, as opposite to the scanned forms handled by VisualDocumentNER. PdfToForm is complementary to VisualDocumentNER.
  
+</div><div class="h3-box" markdown="1">
 
 #### Improvements
 
 * Support for multi-frame dicom has been added.
 * Added the missing load()​ method in ImageToTextV2.
 
- 
+</div><div class="h3-box" markdown="1">
 
 #### New Notebooks
 
diff --git a/docs/en/spark_ocr_versions/release_notes_4_1_0.md b/docs/en/spark_ocr_versions/release_notes_4_1_0.md
index 6fb0bc1a11..802896ffe1 100644
--- a/docs/en/spark_ocr_versions/release_notes_4_1_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_4_1_0.md
@@ -20,8 +20,9 @@ Release date: 22-09-2022
 #### Overview
 
 We are glad to announce that Spark OCR 4.1.0 has been released!
-This release comes with new features, enhancements, fixes and more!.
- 
+This release comes with new features, enhancements, fixes and more!
+
+</div><div class="h3-box" markdown="1">
 
 #### New Features
 * DicomSplitter: new annotator that helps to distribute and split Dicom files into multiple frames. It supports multiple strategies, similar to our PdfToImage annotator. It enables parallel processing of different frames and keeps memory utilization bounded. For big datasets, or memory constrained environments, it enables Streaming Mode to process frames 1-by-1, resulting in very low memory requirements.
@@ -50,12 +51,14 @@ This release comes with new features, enhancements, fixes and more!.
 
 
 * New image pre-processing annotators: ImageHomogenizeLight, ImageRemoveBackground, ImageEnhanceContrast, ImageRemoveGlare. For examples on how to use them, and their amazing results check this notebook: [SparkOcrImagePreprocessing.ipynb](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/master/jupyter/SparkOcrImagePreprocessing.ipynb).
- 
+
+</div><div class="h3-box" markdown="1">
 
 #### Improvements
 * VisualDocumentClassifierV2 training has been improved for more efficient memory utilization.
 * Library dependencies have been updated to remove security vulnerabilities.
 
+</div><div class="h3-box" markdown="1">
 
 #### Bug Fixes
 * The infamous "ImportError: No module named resource" bug that was affecting Windows users has been fixed.
diff --git a/docs/en/spark_ocr_versions/release_notes_4_2_0.md b/docs/en/spark_ocr_versions/release_notes_4_2_0.md
index 987b12f5d2..fea9a9e949 100644
--- a/docs/en/spark_ocr_versions/release_notes_4_2_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_4_2_0.md
@@ -17,15 +17,18 @@ sidebar:
 
 Release date: 31-10-2022
 
-
 We are glad to announce that Spark OCR 4.2.0 has been released. This is mostly a compatibility release to ensure compatibility of Spark OCR against Spark NLP 4.2.1, and Spark NLP Healthcare 4.2.1.
 
 #### Improvements
 * Improved memory consumption and performance in the training of Visual NER models.
 
+</div><div class="h3-box" markdown="1">
+
 #### New Features
 * PdfToForm new param: useFullyQualifiedName, added capability to return fully qualified key names.
 
+</div><div class="h3-box" markdown="1">
+
 #### New or Updated Notebooks
 * [SparkOcrProcessMultiplepageScannedPDF.ipynb](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/master/jupyter/SparkOcrProcessMultiplepageScannedPDF.ipynb) has been added to show how to serve a multi-page document processing pipeline.
 * [SparkOcrDigitalFormRecognition.ipynb](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/4.2.0-release-candidate/jupyter/FormRecognition/SparkOcrDigitalFormRecognition.ipynb) has been updated to show utilization of useFullyQualifiedName parameter.
diff --git a/docs/en/spark_ocr_versions/release_notes_4_2_1.md b/docs/en/spark_ocr_versions/release_notes_4_2_1.md
index 821f5d754b..44c55d47ad 100644
--- a/docs/en/spark_ocr_versions/release_notes_4_2_1.md
+++ b/docs/en/spark_ocr_versions/release_notes_4_2_1.md
@@ -19,7 +19,8 @@ Release date: 11-28-2022
 
  
 We're glad to announce that Spark-OCR 4.2.1 has been released! This release is almost completely about LightPipelines.
- 
+
+</div><div class="h3-box" markdown="1">
 
 #### LightPipeline added to Spark-OCR
 Originally introduced by Spark-NLP, this has been one of the most celebrated features by our users. In a nutshell, LightPipelines allow you switching your pipeline from distributed processing to local mode, in a single line of code. Also, results are much easier to post-process as they come in plain Python data structures. 
@@ -31,10 +32,12 @@ Now, LightPipelines are available in Spark-OCR as well! This is an initial imple
 
 Following is a chart comparing performance of different techniques on batches of different page counts: 8, 16, 24, 32, 40, 48, and 80 pages.
 
-![image](/assets/images/ocr/light_pipelines.png)
+![LightPipeline added to Spark-OCR](/assets/images/ocr/light_pipelines.png)
 
 For the 8 pages case, on the left side of the chart, LightPipelines average 1.25s per page vs. 4s per page that were scored by a similar Pytesseract implementation. That makes LightPipelines a great candidate to achieve low latency on small sized batches, while still leveraging parallelism.
 
+</div><div class="h3-box" markdown="1">
+
 #### Korean Support
 You can start using Korean language by just passing the 'KOR' option to ImageToText,
 ```python
@@ -47,9 +50,13 @@ You can start using Korean language by just passing the 'KOR' option to ImageToT
     ocr.setDownloadModelData(True)
 ```
 
+</div><div class="h3-box" markdown="1">
+
 #### Bug Fixes
 * AlabReader has been updated to handle the new structure present in Annotation Lab's exported annotations.
 
+</div><div class="h3-box" markdown="1">
+
 #### New Notebooks
 * Check how to use LightPipelines in this notebook: [SparkOcrLightPipelines.ipynb](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/4.2.1-release-candidate/jupyter/SparkOcrLightPipelines.ipynb)
 
diff --git a/docs/en/spark_ocr_versions/release_notes_4_2_4.md b/docs/en/spark_ocr_versions/release_notes_4_2_4.md
index e04155f1f8..cb579bc525 100644
--- a/docs/en/spark_ocr_versions/release_notes_4_2_4.md
+++ b/docs/en/spark_ocr_versions/release_notes_4_2_4.md
@@ -15,13 +15,17 @@ sidebar:
 
 ## 4.2.4
 
-We are glad to announce that Spark OCR 4.2.4⚡has been released!! 
+We are glad to announce that Spark OCR 4.2.4⚡has been released!
 This release includes new optimized ImageToTextV2 models, more support on annotators in LightPipelines, a new PdfToHocr annotator, enhancements, and more!
 
+</div><div class="h3-box" markdown="1">
+
 #### New Features
 * New annotators supported in LightPipelines: PdfToText and most Image transformations. Check sample notebook for details.
 * Handling of PDFs with broken headers: some PDFs may contain incorrect header information causing the pipelines to fail to process them, now PDF processing annotators support handling these documents.
 
+</div><div class="h3-box" markdown="1">
+
 #### New Annotators
 * New ImageToTextV2 Transformers based OCR annotator, 
   * Intended to become a full replacement of original ImageToTextV2.
@@ -41,6 +45,7 @@ ocr = ImageToTextV2.pretrained("ocr_base_printed_v2_opt", "en", "clinical/ocr")
 
 * New PdfToHocr: this new annotator allows to produce HOCR output from digital PDFs. This is not only useful for integrating into existing annotators that already consume HOCR, but for new pipelines that will be released in the future. Stay tuned for new releases.  
 
+</div><div class="h3-box" markdown="1">
 
 #### New Models
 ocr_base_printed_v2
@@ -48,6 +53,7 @@ ocr_base_handwritten_v2
 ocr_base_printed_v2_opt (quantized version)
 ocr_base_handwritten_v2_opt (quantized version)
 
+</div><div class="h3-box" markdown="1">
 
 #### New Notebooks
 * New supported transformers in LightPipelines in action,
diff --git a/docs/en/spark_ocr_versions/release_notes_4_3_0.md b/docs/en/spark_ocr_versions/release_notes_4_3_0.md
index 2809331fca..edd1a4ea99 100644
--- a/docs/en/spark_ocr_versions/release_notes_4_3_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_4_3_0.md
@@ -19,6 +19,7 @@ Release date: 2023-01-13
 
 We are glad to announce that Spark OCR 4.3.0 has been released!! This big release comes with improvements in Dicom Processing, Visual Question Answering, new Table Extraction annotators, and much more!.
 
+</div><div class="h3-box" markdown="1">
  
 ### New Features
 * PositionFinder now works in LightPipelines.
@@ -42,12 +43,15 @@ Optimized model is smaller(about a half) of the original open-source version, ma
 Two models available: `docvqa_donut_base` and `docvqa_donut_base_opt`(quantized).
 LightPipelines support.
  	
+</div><div class="h3-box" markdown="1">
 
 #### Bug Fixes
 
 * Empty tables now handled properly in ImageCellsToTextTable.
 * Pretrained models for VisualDocumentNerV21 are now accessible.
 
+</div><div class="h3-box" markdown="1">
+
 #### New/updated Notebooks
 
 * [SparkOcrVisualQuestionAnswering.ipynb](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/master/jupyter/SparkOcrVisualQuestionAnswering.ipynb), this notebook shows examples on how to use Donut based visual question answering in Spark-OCR.
diff --git a/docs/en/spark_ocr_versions/release_notes_4_3_1.md b/docs/en/spark_ocr_versions/release_notes_4_3_1.md
index 63c84cb9ff..2c42144de2 100644
--- a/docs/en/spark_ocr_versions/release_notes_4_3_1.md
+++ b/docs/en/spark_ocr_versions/release_notes_4_3_1.md
@@ -19,14 +19,20 @@ Release date: 17-02-2023
 
 We're glad to announce that Visual NLP 😎 4.3.1 has been released.
 
+</div><div class="h3-box" markdown="1">
+
 ### Highlights
 * ImageTextCleaner & ImageTableDetector have improved memory consumption.
 * New Annotators supported in LightPipelines.
 * Table extraction from Digital PDFs pipeline now entirely supported as a LightPipeline.
 
+</div><div class="h3-box" markdown="1">
+
 ### ImageTextCleaner & ImageTableDetector improved memory consumption
 * ImageTextCleaner & ImageTableDetector improved memory consumption: we reduced about 30% the memory consumption for this annotator making it more memory friendly and enabling running on memory constrained environments like Colab.
 
+</div><div class="h3-box" markdown="1">
+
 ### New Annotators supported in LightPipelines
 Now the following annotators are supported in LightPipelines,
 * PdfToHocr,
@@ -35,6 +41,8 @@ Now the following annotators are supported in LightPipelines,
 * ImageScaler,
 * HocrToTextTable,
 
+</div><div class="h3-box" markdown="1">
+
 ### Table extraction from Digital PDFs pipeline now entirely supported as a LightPipeline.
 * Our Table Extraction from digital PDFs pipeline now supports running as a LightPipeline, check the updated notebook: [SparkOCRPdfToTable.ipynb](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/master/jupyter/SparkOCRPdfToTable.ipynb)
 
diff --git a/docs/en/spark_ocr_versions/release_notes_4_3_3.md b/docs/en/spark_ocr_versions/release_notes_4_3_3.md
index c41e0d89ca..b4fc7a1eb3 100644
--- a/docs/en/spark_ocr_versions/release_notes_4_3_3.md
+++ b/docs/en/spark_ocr_versions/release_notes_4_3_3.md
@@ -19,6 +19,8 @@ Release date: 14-03-2023
 
 We're glad to announce that Visual NLP 😎 4.3.3 has been released.
 
+</div><div class="h3-box" markdown="1">
+
 ### Highlights
 * New parameter keepOriginalEncoding in PdfToHocr.
 * New Yolo-based table and form detector. 
@@ -26,33 +28,40 @@ We're glad to announce that Visual NLP 😎 4.3.3 has been released.
 * Fixes in AlabReader
 * Fixes in HocrToTextTable.
 
+</div><div class="h3-box" markdown="1">
+
 #### New parameter keepOriginalEncoding in PdfToHocr
 Now you can choose to make PdfToHocr return an ASCII normalized version of the characters present in the PDF(keepOriginalEncoding=False) or to return the original Unicode character(keepOriginalEncoding=True).
 Source PDF,
-![image](/assets/images/ocr/source.png)
+![keepOriginalEncoding in PdfToHocr](/assets/images/ocr/source.png)
 
 Keeping the encoding,
-![image](/assets/images/ocr/keeping.png)
+![keepOriginalEncoding in PdfToHocr](/assets/images/ocr/keeping.png)
 
 Not keeping it,
-![image](/assets/images/ocr/notkeeping.png)
+![keepOriginalEncoding in PdfToHocr](/assets/images/ocr/notkeeping.png)
 
+</div><div class="h3-box" markdown="1">
 
 #### New Yolo-based Table and Form detector
 This new model allows to distinguish between forms and tables, so you can apply different downstream processing afterwards.
 
-![image](/assets/images/ocr/form_tables.jpg)
+![New Yolo-based Table and Form detector](/assets/images/ocr/form_tables.jpg)
 
 Check a full example of utilization in [this notebook](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/master/jupyter/SparkOcrImageTableAndFormDetection.ipynb).
 
+</div><div class="h3-box" markdown="1">
 
 #### Memory consumption in VisualQuestionAnswering and ImageTableDetector models has been improved
 Memory utilization has been improved to make it more GC friendly. The practical result is that big jobs are more stable, and less likely to get restarted because of exhausting resources.
 
+</div><div class="h3-box" markdown="1">
 
 #### Fixes in AlabReader
 AlabReader has been improved to fix some bugs, and to improve the performance.
 
+</div><div class="h3-box" markdown="1">
+
 #### Fixes in HocrToTextTable
 HocrToTextTable has been improved in order to better handle some corner cases in which the last rows of tables were being missed.
 
diff --git a/docs/en/spark_ocr_versions/release_notes_4_4_0.md b/docs/en/spark_ocr_versions/release_notes_4_4_0.md
index 8447004a9e..c73ce2dcd3 100644
--- a/docs/en/spark_ocr_versions/release_notes_4_4_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_4_4_0.md
@@ -25,6 +25,8 @@ We're glad to announce that Visual NLP 😎 4.4.0 has been released.
 * Improvements
 * New notebooks
 
+</div><div class="h3-box" markdown="1">
+
 #### Pretrained Pipelines for Visual NLP
 We are adding support for pretrained pipelines that will allow to package an entire set of models and transformations into single unit. For starters, this is the list of pretrained pipelines we are releasing today,
 
@@ -36,6 +38,7 @@ We are adding support for pretrained pipelines that will allow to package an ent
 * pdf_printed_transformer_extraction: OCR printed texts contained in PDFs.
 * pdf_handwritten_transformer_extraction: OCR handwritten texts contained in PDFs.
 
+</div><div class="h3-box" markdown="1">
 
 #### Deprecations & Changes
 Some ImageToTextV2 models have been deprecated in favor of new optimized versions that work entirely on the JVM, and also now work in LightPipelines.
@@ -66,10 +69,13 @@ SparkOCRVisualDocumentNERv2.ipynb -> SparkOCRVisualDocumentNer-FormParsing.ipynb
 
 Finally, VisualDocumentNer fine tuning notebooks are currently deprecated until new ones are included on next release.
 
+</div><div class="h3-box" markdown="1">
 
 #### Improvements
 We improved digital PDF data ingestion for documents containing ligature glyphs which are used in some fonts. Ligature glyphs are symbols that may represent two or more characters. For instance, 'f' and 'i' may be represented as one char 'fi'. We improved our transformer to process them correctly.
 
+</div><div class="h3-box" markdown="1">
+
 #### New Notebooks
 [SparkOcrPretrainedPipelines.ipynb](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/master/jupyter/SparkOcrPretrainedPipelines.ipynb): use this notebook to learn how to use the new pretrained pipelines feature added to Visual NLP 4.4.0.
 
diff --git a/docs/en/spark_ocr_versions/release_notes_4_4_1.md b/docs/en/spark_ocr_versions/release_notes_4_4_1.md
index 091d282988..4c653197bf 100644
--- a/docs/en/spark_ocr_versions/release_notes_4_4_1.md
+++ b/docs/en/spark_ocr_versions/release_notes_4_4_1.md
@@ -19,6 +19,8 @@ Release date: 15-05-2023
 
 We are glad to announce that Visual NLP 😎 4.4.1 has been released! This release comes with a number of improvements, bug fixes, new implementations, and more!
 
+</div><div class="h3-box" markdown="1">
+
 ### Highlights
 #### New Features
 * Entirely new implementation for PositionFinder.
@@ -30,10 +32,14 @@ We are glad to announce that Visual NLP 😎 4.4.1 has been released! This relea
 * Improvements in Dicom file processing.
 * Fixes in Python installation process.
 
+</div><div class="h3-box" markdown="1">
+
 #### New notebooks
 
 [SparkOcrLightPipelinesBase64.ipynb](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/master/jupyter/SparkOcrLightPipelinesBase64.ipynb).
 
+</div><div class="h3-box" markdown="1">
+
 #### New Features
 * Entirely new implementation for PositionFinder: PositionFinder has been re-implemented to deal with the limitations of the original version. Many bugs in the previous version simply won't be present in the new implementation. Also, the new version will work with coordinates originated in digital PDFs as well as coordinates originated on images. 
 
@@ -78,6 +84,9 @@ ocr = ImageToTextV2.pretrained("ocr_base_printed_v2_opt", "en", "clinical/ocr")
     .setOutputCol("text") \
     .setTaskParallelism(12)
 ```
+
+</div><div class="h3-box" markdown="1">
+
 #### Fixes
 * Most vulnerabilities at the Java level were removed: 50 out of 52 reported vulnerabilities were removed from the JAR file of Visual NLP. Only 2 medium vulnerabilities remain.
 
@@ -98,6 +107,8 @@ Known Limitations: due to mismatch in column types from different annotators ret
 
 * Fixes in Python installation process: installation process has been improved especially in environments with newer Python versions like Colab.
 
+</div><div class="h3-box" markdown="1">
+
 #### New Notebooks
 SparkOcrLightPipelinesBase64.ipynb: learn how to use the new Base64ToImage Transformer in LightPipelines to feed in memory string buffers to Visual NLP pipelines.
 This release is compatible with Spark NLP for Healthcare 4.4.1, and Spark NLP 4.4.1.
diff --git a/docs/en/spark_ocr_versions/release_notes_4_4_2.md b/docs/en/spark_ocr_versions/release_notes_4_4_2.md
index fb442e8f82..d7a475a1b0 100644
--- a/docs/en/spark_ocr_versions/release_notes_4_4_2.md
+++ b/docs/en/spark_ocr_versions/release_notes_4_4_2.md
@@ -19,6 +19,8 @@ Release date: 30-05-2023
 
 We are glad to announce that Visual NLP 😎 4.4.2 has been released. This is a small release with mostly bug fixes and minor improvements.
 
+</div><div class="h3-box" markdown="1">
+
 #### Fixes
 * ImageTextDetectorV2 initialization bug happening in some cluster environments is now fixed.
 * PdfToText and PdfToHocr now return document dimensions using the same data type(integer).
@@ -29,9 +31,9 @@ New Features
 + Bounding boxes spawning multiple lines are now supported in PositionFinder!
 
 original:
-![image](/assets/images/ocr/position_finder_1.png)
+![4.4.2](/assets/images/ocr/position_finder_1.png)
 masked:
-![image](/assets/images/ocr/position_finder_2.png)
+![4.4.2](/assets/images/ocr/position_finder_2.png)
 
 Here for "Lockheed Martin" PositionFinder will return two bounding boxes. Remember that you can still link the two bounding boxes to the original entity by using the 'chunk index'.
 
diff --git a/docs/en/spark_ocr_versions/release_notes_4_4_3.md b/docs/en/spark_ocr_versions/release_notes_4_4_3.md
index ccfe43a80d..665800c304 100644
--- a/docs/en/spark_ocr_versions/release_notes_4_4_3.md
+++ b/docs/en/spark_ocr_versions/release_notes_4_4_3.md
@@ -19,6 +19,8 @@ Release date: 33-06-2023
 
 We are glad to announce that Visual NLP 😎 4.4.3 has been released. 
 
+</div><div class="h3-box" markdown="1">
+
 #### Highlights
 
 In line with our unwavering dedication to delivering top-notch products, Visual NLP is backed by our steadfast commitment to quality and stability. We have meticulously addressed various issues and introduced enhancements to ensure a seamless user experience. By diligently resolving bugs, improving error handling, and expanding functionality, we prioritize the reliability and robustness of Visual NLP.
diff --git a/docs/en/spark_ocr_versions/release_notes_4_4_4.md b/docs/en/spark_ocr_versions/release_notes_4_4_4.md
index 881e87b8fa..93ce588ce1 100644
--- a/docs/en/spark_ocr_versions/release_notes_4_4_4.md
+++ b/docs/en/spark_ocr_versions/release_notes_4_4_4.md
@@ -41,6 +41,8 @@ Among other things, this helps the process of matching entities to coordinates i
 
 * PositionFinder has an improved matching strategy to map entities to coordinates that prevents entities from remaining unmapped in many situations. Also, error reporting has been improved, making it clear in the logs when for some reason an entity couldn't be located in the document, and coordinates were not returned.
 
+</div><div class="h3-box" markdown="1">
+
 #### Bug Fixes & Improvements
 * Serialization problems in ImageDrawRegions and ImageDrawAnnotations were fixed.
 * Some dependencies have been upgraded to newer versions to maintain .
diff --git a/docs/en/spark_ocr_versions/release_notes_5_0_0.md b/docs/en/spark_ocr_versions/release_notes_5_0_0.md
index bacdb7c03c..0cf317c94c 100644
--- a/docs/en/spark_ocr_versions/release_notes_5_0_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_5_0_0.md
@@ -20,6 +20,8 @@ Release date: 21-08-2023
 We are glad to announce that Visual NLP 😎 5.0.0 has been released! 
 This release comes with new models, bug fixes and more!
 
+</div><div class="h3-box" markdown="1">
+
 #### New Models
 * New dit_base_finetuned_rvlcdip_opt: Dit based Visual Document Classification model. This is an optimized version of previous dit_base_finetuned_rvlcdip model. It has a reduced model size of 80MB(vs. 304 of original model), which reduces the memory footprint, also memory management within the model itself has been improved. It offers a speedup of 1.54x compared to the original implementation. The impact in accuracy is minimal, it achieves an accuracy of 91.55% over RVL-CDIP dataset compared to 91.83% of the original model.
 Setting up the model is straightforward,
@@ -56,16 +58,21 @@ doc_class = VisualDocumentClassifierLilt() \
 ```
 Use [this notebook](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/master/jupyter/SparkOCRVisualDocumentClassifierLiLT.ipynb) as a reference.
 
+</div><div class="h3-box" markdown="1">
 
 #### New Annotators
 * New DicomToPdf and DicomUpdatePdf annotators: the new annotators now make it possible to extract and update encapsulated PDF files within DICOM documents. This opens up opportunities to building de-identification pipelines for the purpose of anonymizing PDF documents that have been encapsulated(embedded) into Dicom files.
 
+</div><div class="h3-box" markdown="1">
+
 #### Bug Fixes
 * ImageDrawAnnotations serialization issues were solved.
 * FormRelationExtraction is now compatible with the new Lilt Visual Ner models.
 * Pipeline serialization issues in Databricks affecting annotators like ImageHandwrittenDetector have been solved.
 * Pillow related errors in Colab setup have been fixed. 
 
+</div><div class="h3-box" markdown="1">
+
 #### New Notebooks
 * [VisualDocumentClassifierTraining](https://github.com/JohnSnowLabs/spark-ocr-workshop/tree/master/jupyter/VisualDocumentClassifierTraining), notebooks for Visual Documents Classifier fine tuning have been updated to use the new Lilt based models.
 * [SparkOcrDeidentificationDicomWithEncapsulatedPDF.ipynb](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/master/jupyter/Dicom/SparkOcrDeidentificationDicomWithEncapsulatedPDF.ipynb), learn how to use the new DicomToPdf and DicomUpdatePdf.
diff --git a/docs/en/spark_ocr_versions/release_notes_5_0_1.md b/docs/en/spark_ocr_versions/release_notes_5_0_1.md
index 22ade0e42f..b59791903b 100644
--- a/docs/en/spark_ocr_versions/release_notes_5_0_1.md
+++ b/docs/en/spark_ocr_versions/release_notes_5_0_1.md
@@ -32,7 +32,7 @@ ImageTextDetector \
 .setScoreThreshold(0.5)
 ```
 
-![image](/assets/images/ocr/image_text_detector_dit.png)
+![5.0.1](/assets/images/ocr/image_text_detector_dit.png)
 
 It is currently the best performing model at the FUNSD dataset, achieving an accuracy of 94% vs Craft detector which achieved 78.7%, and is recommended for De-identification and Text Extraction pipelines.
 
diff --git a/docs/en/spark_ocr_versions/release_notes_5_0_2.md b/docs/en/spark_ocr_versions/release_notes_5_0_2.md
index 4a535d6f18..14af370a29 100644
--- a/docs/en/spark_ocr_versions/release_notes_5_0_2.md
+++ b/docs/en/spark_ocr_versions/release_notes_5_0_2.md
@@ -37,14 +37,10 @@ This is a small compatibility release to ensure the product runs smoothly on Goo
 
 This release is compatible with ```Spark NLP 5.1.1``` and Spark NLP for``` Healthcare 5.1.1```
 
-
-
-
-
 </div><div class="prev_ver h3-box" markdown="1">
 
 ## Previous versions
 
 </div>
 
-{%- include docs-sparckocr-pagination.html -%}
+{%- include docs-sparckocr-pagination.html -%}
\ No newline at end of file
diff --git a/docs/en/spark_ocr_versions/release_notes_5_1_0.md b/docs/en/spark_ocr_versions/release_notes_5_1_0.md
index 1364b01d16..74ad82db3a 100644
--- a/docs/en/spark_ocr_versions/release_notes_5_1_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_5_1_0.md
@@ -18,14 +18,14 @@ sidebar:
 Release date: 17-11-2023
 
 
-**We are glad to announce that Visual NLP 5.1.0 has been released! This release comes with new models, annotators, bug fixes, and more!.📢📢📢**
+**We are glad to announce that Visual NLP 5.1.0 has been released! This release comes with new models, annotators, bug fixes, and more!📢📢📢**
 
 **New Models &  Annotators**
 * VisualQuestionAnsweringPix2Struct: we are adding a new Visual Question Answering(VQA) checkpoint for Pix2Struct. Document VQA is the task of answering questions about documents, in which visual clues are important in the answer.
 The practical impact of this type of models is that you can create "data extractors" for your own particular use case without fine-tuning on your data. So you can ask questions about tables, or forms or other structures in which the visual information is relevant, in a zero-shot manner.
 
 We started our journey with Donut-like models, which were great in many different tasks. Check [this code and example](https://nlp.johnsnowlabs.com/2023/01/17/docvqa_donut_base_en_3_2.html), and [this webinar](https://www.johnsnowlabs.com/watch-zero-shot-visual-question-answering/), in case you missed it.
-![image](/assets/images/ocr/pix2struct_sample.png)
+![5.1.0](/assets/images/ocr/pix2struct_sample.png)
 
 ```
 |[What's the estimated population in poverty of Lawrence? ->  5,696, What's the population of Stoddard? ->  26,000, What is the page number of the document? ->  6, What is the date in the document? ->  January, 1970]|
@@ -36,7 +36,7 @@ Now, we're taking one step further and integrating Pix2Struct which, when compar
 Check [this notebook](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/master/jupyter/SparkOcrVisualQuestionAnsweringJsl.ipynb) with examples on how to use it.
 
 * DocumentLayoutAnalyzer: document layout analysis is a fundamental task in Visual NLP, it is the task of detecting sections in a document. Typical examples for these sections are: text, title, list, table, or figure.
-![image](/assets/images/ocr/dit-layout-sample.png)
+![5.1.0](/assets/images/ocr/dit-layout-sample.png)
 
 
  
diff --git a/docs/en/spark_ocr_versions/release_notes_5_1_2.md b/docs/en/spark_ocr_versions/release_notes_5_1_2.md
index 846f73cef3..a16649120e 100644
--- a/docs/en/spark_ocr_versions/release_notes_5_1_2.md
+++ b/docs/en/spark_ocr_versions/release_notes_5_1_2.md
@@ -18,9 +18,9 @@ sidebar:
 Release date: 03-01-2024
 
 
- ## Visual NLP 5.1.2 Release Notes 🕶️
-
+## Visual NLP 5.1.2 Release Notes 🕶️
 
+</div><div class="h3-box" markdown="1">
 
 **We are glad to announce that Visual NLP 5.1.2 has been released!TThis release comes with faster than ever OCR models, improved Table Extraction pipelines, bug fixes, and more! 📢📢📢**
 
@@ -31,10 +31,14 @@ Release date: 03-01-2024
 + Other Changes.
 + Bug fixes.
 
+</div><div class="h3-box" markdown="1">
+
 ## New optimized OCR checkpoints with up to 5x speed ups and GPU support 🚀
 ImageToTextV2, is our Transformer-based OCR model which delivers SOTA accuracy across different pipelines like Text Extraction(OCR), Table Extraction, and Deidentification. </br>
 We've added new checkpoints together with more options to choose which optimizations to apply.
 
+</div><div class="h3-box" markdown="1">
+
 ### New checkpoints for ImageToTextV2 📍
 All previous checkpoints have been updated to work with the latest optimizations, and in addition these 4 new checkpoints have been added,
 
@@ -45,21 +49,25 @@ All previous checkpoints have been updated to work with the latest optimizations
 
 These 4 checkpoints are more accurate than their 'base' counterparts. We are releasing metrics for the 'base' checkpoints today, and a full chart including these checkpoints will be presented in a blogpost to be released soon.
 
+</div><div class="h3-box" markdown="1">
+
 ### New options for ImageToTextV2 ⚡️
 ImageToTextV2 now supports the following configurations:
 * setUseCaching(Boolean): whether or not to use caching during processing.
 * setBatchSize(Integer): the batch size dictates the size of the groups that are processes internally at a single time by the model, typically used when setUseGPU() is set to true. 
 
+</div><div class="h3-box" markdown="1">
+
 ### Choosing the best checkpoint for your problem 💥
 We put together this grid reflecting performance and accuracy metrics to help you choose the most appropriate checkpoint for your use case.
 * Accuracy
 
-![image](/assets/images/ocr/acc.png)
+![5.1.2](/assets/images/ocr/acc.png)
 
 
 * Performance
 
-![image](/assets/images/ocr/perf.png)
+![5.1.2](/assets/images/ocr/perf.png)
 
 Note:
 * CER: character error rate.
@@ -67,6 +75,8 @@ Note:
 * The CPU cluster is a 30 node cluster of 64 DBU/h, and the GPU cluster is a 10 node cluster, of 15 DBU/h.
 * Compared to previous releases, the optimizations introduced in this release yield a speed up of almost 5X, and a cost reduction of more than 4 times, if GPU is used.
 
+</div><div class="h3-box" markdown="1">
+
 ## New improved Table Extraction Pipeline with improved Cell Detection stage. 🔥
 Starting in this release, our HocrToTextTable annotator can receive information related to cells regions to improve the quality of results in Table Extraction tasks. This is particularly useful for cases in which cells are multi-line, or for borderless tables.  </br>
 This is what a pipeline would look like,
@@ -104,6 +114,7 @@ For a complete, end-to-end example we encourage you to check the sample notebook
 
 [SparkOcrImageTableRecognitionWHOCR.ipynb](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/master/jupyter/SparkOcrImageTableRecognitionWHOCR.ipynb)
 
+</div><div class="h3-box" markdown="1">
 
 ## Other changes 🎯
 * Dicom private tags in metadata now can be removed in DicomMetadataDeidentifier: calling  setRemovePrivateTags(true) will cause the tags marked as private to be removed in the output Dicom document.
@@ -112,42 +123,21 @@ For a complete, end-to-end example we encourage you to check the sample notebook
 * start() function now supports the configuration of GPU through the boolean `use_gpu` parameter.
 * Faster(20%), and smaller footprint, `docvqa_pix2struct_jsl_opt` Visual Question Answering checkpoint.
 
+</div><div class="h3-box" markdown="1">
+
 ## Bug Fixes 🪲
 * ImageSplitRegions does not work after Table Detector.
 * VisualDocumentNerLilt output has been fixed to include entire tokens instead of pieces.
 * Null Regions in HocrToTextTable are handled properly.
 * display_tables can now handle empty tables better.
 * Vulnerabilities in Python dependencies.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
 * This release is compatible with ```Spark NLP 5.2.0``` and Spark NLP for``` Healthcare 5.1.1```
 
 
-
-
-
 </div><div class="prev_ver h3-box" markdown="1">
 
 ## Previous versions
 
 </div>
 
-{%- include docs-sparckocr-pagination.html -%}
+{%- include docs-sparckocr-pagination.html -%}
\ No newline at end of file
diff --git a/docs/en/spark_ocr_versions/release_notes_5_2_0.md b/docs/en/spark_ocr_versions/release_notes_5_2_0.md
index 5858041ce6..35ba628a56 100644
--- a/docs/en/spark_ocr_versions/release_notes_5_2_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_5_2_0.md
@@ -18,12 +18,11 @@ sidebar:
 Release date: 23-02-2024
 
 
- ## Visual NLP 5.2.0 Release Notes 🕶️
-
-
+## Visual NLP 5.2.0 Release Notes 🕶️
 
 **We are glad to announce that Visual NLP 5.2.0 has been released. This release comes with new models, bug fixes, blog posts, and more!! 📢📢📢**
 
+</div><div class="h3-box" markdown="1">
 
 ## Highlights 🔴
 + New Chart-To-Text dePlot based models.
@@ -36,55 +35,67 @@ Release date: 23-02-2024
 + Bug Fixes
 + Other Changes
 
+</div><div class="h3-box" markdown="1">
+
 ## New Chart-To-Text dePlot based models 📈
 Chart To Text is the task of converting an image chart into a serialized textual version representation of the chart. To understand this, consider the following example,
 
-![image](/assets/images/ocr/chart.png)
+![5.2.0](/assets/images/ocr/chart.png)
 
 Maps to the following text based representation,
 
-![image](/assets/images/ocr/extracted_graph.png)
+![5.2.0](/assets/images/ocr/extracted_graph.png)
 
 For an end-to-end example, please check [this notebook](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/master/jupyter/SparkOcrChartToTextTable.ipynb).
 
+</div><div class="h3-box" markdown="1">
+
 ## Support for Confidence Scores in Visual Question Answering Models. 📍
 Now, VisualQuestionAnswering models support confidence scores. The output schema for VisualQuestionAnswering models has been updated to include questions, answers and confidence scores. To enable confidence scores in the output of these models you should call `setConfidenceScore(true)`. For example,
 
 
-![image](/assets/images/ocr/new_schema_vqa.png)
+![5.2.0](/assets/images/ocr/new_schema_vqa.png)
 
 shows the schema and sample output for the case of two questions, with their corresponding answers and confidence scores.
 
+</div><div class="h3-box" markdown="1">
+
 ## Improved stability and new metrics for ImageToTextV2 models. ⚡️
 ImageToTextV2, our Transformer-based OCR has been improved, and extensively stress tested for stability and reliability.
 These are the latest metrics for accuracy and runtime performance for all checkpoints,
 
-![image](/assets/images/ocr/ocr_table.png)
+![5.2.0](/assets/images/ocr/ocr_table.png)
 
 A key takeaway from this chart is the following: _The [Dbu/h] is four times higher for CPU compared to GPU, with no variance in accuracy. Utilizing GPU can achieve identical outcomes at one-fourth of the cost. GPU is your friend!_
 
+</div><div class="h3-box" markdown="1">
+
 ## New Blog Post in ImageToTextV2 models. 💥
 Want to learn about the best practices to scale out your OCR pipelines?. Read the full article [here](https://medium.com/john-snow-labs/unleashing-the-power-of-high-throughput-inference-with-tr-ocr-1f0fa3bc46c6).
 
+</div><div class="h3-box" markdown="1">
 
 ## Docker image for Visual NLP. 🔥
 For users that require running inside a container we have created the following [instructions](https://github.com/JohnSnowLabs/spark-ocr-workshop/tree/master/docker/visual-ner) and [sample notebook](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/master/docker/visual-ner/content/VisualNlp.ipynb).
 
+</div><div class="h3-box" markdown="1">
+
 ## New Pretrained pipeline `basic_table_extractor`
 
 This is a complete Table Extraction Pipeline. Following, it's a basic example of how to call this pipeline,
-![image](/assets/images/ocr/basic_table_extractor.png)
+![5.2.0](/assets/images/ocr/basic_table_extractor.png)
 
 And you should also check the full example in [this notebook](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/master/webinars/zs_table_processing/TableExtractionBasics.ipynb).
 
+</div><div class="h3-box" markdown="1">
+
 ## Spark 3.5 support 🎯
 We extended support to Apache Spark 3.5. All tests were run using Spark 3.5 and Python 3.10.
 
+</div><div class="h3-box" markdown="1">
+
 ## Other Changes
 * Pix2struct models now support caching, both docvqa_pix2struct_jsl and docvqa_pix2struct_jsl_opt pix2struct based checkpoints now support caching, which is enabled by default.
-
-
-
 * This release is compatible with ```Spark NLP 5.2.2``` and Spark NLP for``` Healthcare 5.2.1```
 
 </div><div class="prev_ver h3-box" markdown="1">
@@ -93,4 +104,4 @@ We extended support to Apache Spark 3.5. All tests were run using Spark 3.5 and
 
 </div>
 
-{%- include docs-sparckocr-pagination.html -%}
+{%- include docs-sparckocr-pagination.html -%}
\ No newline at end of file
diff --git a/docs/en/spark_ocr_versions/release_notes_5_3_0.md b/docs/en/spark_ocr_versions/release_notes_5_3_0.md
index 69a0981bf9..220f2eaeb4 100644
--- a/docs/en/spark_ocr_versions/release_notes_5_3_0.md
+++ b/docs/en/spark_ocr_versions/release_notes_5_3_0.md
@@ -18,11 +18,11 @@ sidebar:
 Release date: 20-03-2024
 
 
- ## Visual NLP 5.3.0 Release Notes 🕶️
-
+## Visual NLP 5.3.0 Release Notes 🕶️
 
 **We are glad to announce that Visual NLP 5.3.0 has been released!. This is a small compatibility release, to ensure proper inter-operability against Spark NLP and Spark NLP for Healthcare, plus some LightPipelines features.!! 📢📢📢**
 
+</div><div class="h3-box" markdown="1">
 
 ## Highlights 🔴
 
@@ -30,6 +30,8 @@ Release date: 20-03-2024
 + New Base64ToBinary annotator.
 + Positions output in ImageToText annotator has been fixed for LightPipelines.
 
+</div><div class="h3-box" markdown="1">
+
 ## New Base64ToBinary Annotator
 This new annotator, as its name suggests, serves the purpose of converting in-memory base64 encoded buffers into binary format. This is useful, among other things, for the use case in which your data is not stored in disk, and it can be really helpful in combination with PdfToText, or PdfToImage, e.g.,
 
diff --git a/docs/en/spark_ocr_versions/release_notes_5_3_1.md b/docs/en/spark_ocr_versions/release_notes_5_3_1.md
index 4686bf68fe..f963d75678 100644
--- a/docs/en/spark_ocr_versions/release_notes_5_3_1.md
+++ b/docs/en/spark_ocr_versions/release_notes_5_3_1.md
@@ -18,11 +18,11 @@ sidebar:
 Release date: 11-04-2024
 
 
- ## Visual NLP 5.3.1 Release Notes 🕶️
-
+## Visual NLP 5.3.1 Release Notes 🕶️
 
 **we're glad to announce that Visual NLP 5.3.1 has been released. New models, notebooks, bug fixes and more!!! 📢📢📢**
 
+</div><div class="h3-box" markdown="1">
 
 ## Highlights 🔴
 
@@ -36,6 +36,8 @@ Release date: 11-04-2024
 + New notebooks & updates.
 + Bug Fixes.
 
+</div><div class="h3-box" markdown="1">
+
 ## Improved table extraction capabilities in HocrToTextTable
 Many issues related to column detection in our Table Extraction pipelines are addressed in this release, compared to previous Visual NLP version the metrics have improved. Table below shows F1-score(CAR or Cell Adjacency Relationship) performances on ICDAR 19 Track B dataset for different IoU values of our two versions in comparison with [other results](https://paperswithcode.com/paper/multi-type-td-tsr-extracting-tables-from/review/).
 
@@ -47,6 +49,7 @@ Many issues related to column detection in our Table Extraction pipelines are ad
 | Visual NLP 5.3.0  | 0.463  | 0.420 | 0.355 | 0.143 |
 | Visual NLP 5.3.1  | 0.509  | **0.477** | **0.403** | **0.162** |
 
+</div><div class="h3-box" markdown="1">
 
 ### Improvements in LightPipelines
 * ImageSplitRegions, ImageToTextV2, ImageTextDetectorCraft are now supported to be used with LightPipelines.
@@ -80,6 +83,8 @@ result = lp.fromString(base64_pdf)
 [Full Example here.](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/master/jupyter/SparkOcrLightPipelinesBase64.ipynb)
 * new `LightPipeline.fromBinary()` method that allows the usage of in-memory binary buffers as inputs to Visual NLP pipelines.
 
+</div><div class="h3-box" markdown="1">\
+
 ### Confidence scores in ImageToTextV2
 You can now enable confidence scores in ImageToTextV2 like this,
 
@@ -88,16 +93,18 @@ ocr = ImageToTextV2.pretrained("ocr_base_printed_v2_opt", "en", "clinical/ocr")
     .setIncludeConfidence(True)
 ```
 
-![image](/assets/images/ocr/confidence_scores.png)
+![5.3.1](/assets/images/ocr/confidence_scores.png)
 
 Check this [updated notebook](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/master/jupyter/TextRecognition/SparkOcrImageToTextV2.ipynb) for an end-to-end example.
 
-
+</div><div class="h3-box" markdown="1">
 
 ### New HocrMerger annotator
 HocrMerger is a new annotator whose purpose is to allow merging two streams of HOCRs texts into a single unified HOCR representation.
 This allows mixing object detection models with text to create a unified document representation that can be fed to other downstream models like Visual NER. The new Checkbox detection pipeline uses this approach.
 
+</div><div class="h3-box" markdown="1">
+
 ###  Checkbox detection in Visual NER.
 A new Checkbox detection model has been added to Visual NLP 5.3.1!. With this model you can detect checkboxes in documents and obtain an HOCR representation of them. This representation, along with the other elements in page can be fed to other models like Visual NER.
 
@@ -112,8 +119,9 @@ check_box_detector = ImageCheckBoxDetector \
 ```
 
 In this case we are receiving an image as input, and returning a HOCR representation with labels 'Yes', and 'No' to represent whether the checkbox is marked or not. You can see this idea directly in the following picture,
-![image](/assets/images/ocr/checkboxes.png)
+![5.3.1](/assets/images/ocr/checkboxes.png)
 
+</div><div class="h3-box" markdown="1">
 
 ### New Document Clustering Pipeline using Vit Embeddings.
 Now we can use Vit Embeddings to create document representations for clustering.
@@ -130,6 +138,8 @@ embeddings = VitImageEmbeddings \
 ```
 For an end-to-end example, please check [this notebook](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/master/jupyter/Clustering/VisualDocumentClustering.ipynb).
 
+</div><div class="h3-box" markdown="1">
+
 ### Enhancements to color options in ImageDrawRegions.
 ImageDrawRegions is the annotator used for rendering regions into images so we can visualize results from different models, like, for example, Text Detection. The setColorMap method, is used to set the colors of bounding boxes drawn at top of images, the new behavior is as follows,
  
@@ -137,6 +147,8 @@ ImageDrawRegions is the annotator used for rendering regions into images so we c
   * setColorMap: when called with a dictionary in the form: {label -> color}, it will apply a different color to each label. If a key is missing, it will pick a random value.
   * setColorMap is not called: random colors will be picked for each label, each time you call transform() a new set of colors will be selected.
 
+</div><div class="h3-box" markdown="1">
+
 ### New notebooks & updates
 + New [notebook](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/master/jupyter/SparkOCRInfographicsVisualQuestionAnswering.ipynb) for Visual Question Answering on Infographics!
 
@@ -144,6 +156,8 @@ ImageDrawRegions is the annotator used for rendering regions into images so we c
 
 + New [notebook](https://github.com/JohnSnowLabs/spark-ocr-workshop/blob/master/jupyter/SparkOCRHandwrittenAndSignatureDetection.ipynb) on Signature and Handwritten text detection.
 
+</div><div class="h3-box" markdown="1">
+
 ### Bug Fixes
 + PdfToImage resetting page information when used in the same pipeline as PdfToText: When the sequence {PdfToText, PdfToImage} was used the original pages computed at PdfToText where resetted to zero by PdfToImage.
 
@@ -154,4 +168,4 @@ ImageDrawRegions is the annotator used for rendering regions into images so we c
 
 </div>
 
-{%- include docs-sparckocr-pagination.html -%}
+{%- include docs-sparckocr-pagination.html -%}
\ No newline at end of file
diff --git a/docs/en/version_compatibility.md b/docs/en/version_compatibility.md
index d9748a7968..39a119e2e7 100644
--- a/docs/en/version_compatibility.md
+++ b/docs/en/version_compatibility.md
@@ -13,7 +13,7 @@ sidebar:
 
 <div class="h3-box" markdown="1">
 
-
+{:.table-model-big}
 | Healthcare NLP | Spark OCR | Spark NLP |
 |----------------|-----------|-----------|
 | 3.0.0          | 3.4.0     | 3.0.0     |