Offline Bundles 2023.4.5

Each offline bundle contains base images used by multiple ML Packages. To make sure which bundle to download for a certain model version, check their compatibility.

Note:

In a scenario where you are using two Automation Suite offline environments - one for development and the other for production, it is important to mirror version consistency. For example, if you are training models on version 2024.10.0 in a development cluster and then move these models to a production cluster, you need to install bundle version 2024.10.0. This is an exceptional use case. As a best practice, always use the latest version of your models for the best results.

Important:

To ensure the compatibility of the model you want to install, check the table below.

Model version	Platform version
	2.2510	2024.10	2023.10	2023.4	2022.10	2022.4
2.2510	✅	❌	❌	❌	❌	❌
2024.10	✅	✅	❌	❌	❌	❌
2023.10	❌	✅	✅	❌	❌	❌
2023.4	❌	❌	✅	✅	❌	❌
2022.10	❌	❌*	❌*	✅	✅	❌
2022.4	❌	❌*	❌*	❌*	❌*	✅

***** - Models still work if you upgrade the platform, but they are not actively maintained in this platform version and security issues are no longer fixed. We recommend updating to a newer model to maintain a safe security posture in your environments.

Offline Bundle for OCR for Chinese, Japanese, Korean

The CJKOCR offline bundle needs to be installed for the ML Packages in the table below. This bundle can be used only on a CPU VM.

If you want to enable Chinese, Japanese, Korean OCR in an offline environment, you also need to follow these steps.

ML Packages	Model version	Metadata
OCR for Chinese, Japanese, Korean	23.4.5	N/A

Offline Bundle for Out-of-the-box Pre-trained ML Packages

The Out-of-the-box Pre-trained ML Packages offline bundle needs to be installed for the ML Packages in the table below.

ML Packages	Model version	Metadata
DocumentUnderstanding	23.4.5	Here
Invoices	23.4.5	Here
InvoicesAustralia	23.4.5	Here
InvoicesIndia	23.4.5	Here
InvoicesJapan	23.4.5	Here
InvocesChina	23.4.5	Here
Receipts	23.4.5	Here
Certificate of Incorporation/ Good Standing	23.4.5	Here
Certificates of Origin	23.4.5	Here
Children Product Certificate	23.4.5	Here
CMS1500	23.4.5	Here
EU Declaration of Conformity	23.4.5	Here
Invoices Shipping	23.4.5	Here
Pay slips	23.4.5	Here
PurchaseOrders	23.4.5	Here
UtilityBills	23.4.5	Here
IDCards	23.4.5	Here
Passports	23.4.5	Here
RemittanceAdvices	23.4.5	Here
BillsOfLading	23.4.5	Here
W2	23.4.5	Here
W9	23.4.5	Here
ACORD125	23.4.5	Here
I9	23.4.5	Here
990	23.4.5	Here
4506T	23.4.5	Here
FM1003	23.4.5	Here
ACORD25	23.4.5	Here
ACORD131	23.4.5	Here
ACORD126	23.4.5	Here
ACORD140	23.4.5	Here
1040	23.4.5	Here
Checks	23.4.5	Here
Bank Statements	23.4.5	Here
Financial statements	23.4.5	Here
Packing Lists	23.4.5	Here
Vehicle Titles	23.4.5	Here

Offline Bundle for Document Classifier

The Document Classifier offline bundle needs to be installed for the ML Packages in the table below.

ML Packages	Model version	Metadata
DocumentClassifier	23.4.5	Here

Offline Bundle Dulv4 (latest Version)

The dulv4 offline bundle needs to be installed only if you want to use ML Packages from the 2022.4.0 enterprise release with AI Center version 2023.4.

For the metadata links, check out the tables in the 2022.4 Document Understanding User Guide.

Offline Bundle Dulv3

The dulv3 offline bundle needs to be installed only if you want to use ML Packages from the 2021.10 enterprise release with AI Center version 2023.4.

For the metadata links, check out the tables in the 2021.10 Document Understanding User Guide.