IBM Automation Document Processing system requirements when disabling deep learning object detection for fixed-format documents in version 24.0.0

Detailed System Requirements

Abstract

Deep learning object detection is an advanced capability that generalizes the annotations from your training documents and dynamically applies them when possible. If your documents have a fixed format and the fields are located in the same places, you don't typically need this capability. When deep learning object detection is disabled, IBM Automation Document Processing extracts the fields from the same positions where they were annotated in the page. This solution works well on those fixed-format documents such as tax forms. If your documents have a dynamic format or sections with variable length, such as invoices, enabling deep learning object detection might yield better accuracy.

If you disable deep learning object detection, the performance is improved for document processing and data extraction training.

Content

You can use the following configuration to disable the deep-learning-object-detection container when you deploy IBM Automation Document Processing, starting with version 24.0.0.

ca_configuration:
  ocrextraction:
    deep_learning_object_detection:
      enabled: false

Attention: The values in the hardware requirements tables were derived under specific operating and environment conditions. The information is accurate under these conditions, but results that are obtained in your operating environments might vary significantly. Therefore, IBM cannot provide any representations, assurances, guarantees, or warranties as to the performance of the profiles in your environment.

Small profile recommendations for Document Processing Engine components:

ca_configuration:
  global:
    deployment_profile_size: "small"

Component	CPU Request (m)	CPU Limit (m)	Memory Request (Mi)	Memory Limit (Mi)	Number of Replicas	Pods are licensed for production and nonproduction	Ephemeral Storage Limit (Mi)
OCR Extraction	200	1000	1024	2560	5	Yes	3072
Classify Process	200	500	400	2048	1	Yes	3072
Processing Extraction	500	1000	1024	3584	3	Yes	3072
Natural Language Extractor	200	500	600	1440	2	Yes	3072
Postprocessing	200	1000	400	1229	1	No	3072
Setup	200	1000	600	2048	2	No	3072
Backend	200	1000	400	2048	2	No	4608
Webhook	200	300	400	500	1	No	1024
RabbitMQ	100	1000	100	1024	2	No	3072
WDU Extraction (technology preview)	300	1000	500	1024	1	No	3072
WDU Runtime (technology preview)	200	4000	1024	8192	1	No	4096

Medium profile recommendations for Document Processing Engine components:

ca_configuration:
  global:
    deployment_profile_size: "medium"

Component	CPU Request (m)	CPU Limit (m)	Memory Request (Mi)	Memory Limit (Mi)	Number of Replicas	Pods are licensed for production and nonproduction
OCR Extraction	200	1000	1024	2560	8	Yes
Classify Process	200	500	400	2048	2	Yes
Processing Extraction	500	1000	1024	3584	3	Yes
Natural Language Extractor	200	500	600	1440	2	Yes
Postprocessing	200	1000	400	1229	2	No
Setup	200	1000	600	2048	4	No
Backend	200	1000	400	2048	4	No
Webhook	200	300	400	500	2	No
RabbitMQ	100	1000	100	1024	3	No
WDU Extraction (technology preview)	300	1000	500	1024	1	No
WDU Runtime (technology preview)	200	4000	1024	8192	1	No

Large profile recommendations for Document Processing Engine components:

ca_configuration:
  global:
    deployment_profile_size: "large"

Component	CPU Request (m)	CPU Limit (m)	Memory Request (Mi)	Memory Limit (Mi)	Number of Replicas	Pods are licensed for production and nonproduction
OCR Extraction	200	1000	1024	2560	13	Yes
Classify Process	200	500	400	2048	3	Yes
Processing Extraction	500	1000	1024	3584	6	Yes
Natural Language Extractor	200	500	600	1440	3	Yes
Postprocessing	200	1000	400	1229	2	No
Setup	200	1000	600	2048	6	No
Backend	200	1000	400	2048	6	No
Webhook	200	300	400	500	3	No
RabbitMQ	100	1000	100	1024	3	No
WDU Extraction (technology preview)	300	1000	500	1024	1	No
WDU Runtime (technology preview)	200	4000	1024	8192	1	No

[{"Type":"MASTER","Line of Business":{"code":"LOB10","label":"Data and AI"},"Business Unit":{"code":"BU048","label":"IBM Software"},"Product":{"code":"SS2JQC","label":"IBM Cloud Pak for Automation"},"ARM Category":[{"code":"a8m3p000000hAKPAA2","label":"Operate-\u003EADP Install\\Upgrade\\Setup"}],"Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"All Versions"}]

Was this topic helpful?

Document Information

More support for:
IBM Cloud Pak for Automation

Component:
Operate->ADP Install\Upgrade\Setup

Software version:
All Versions

Document number:
7151706

Modified date:
28 June 2024

UID

ibm17151706

IBM Support

Tips