Skip to Content
Kontakt os
Chat nu Chat offline
Få direkte hjælp og chat med en SAP-repræsentant.
Kontakt os
Send os en e-mail med kommentarer, spørgsmål eller feedback.

SAP Leonardo Machine Learning Foundation Features

Our enterprise-grade, cloud-based platform for machine learning enables simple consumption and tight integration with SAP software.

Custom Machine Learning Models

Deploy custom machine learning models or customise models with your own data.
Previous Next

Bring your own model

If you have your own TensorFlow model you can easily import and deploy it on the SAP Leonardo Machine Learning Foundation. Once you have done this, you can publish your model as a web service and serve it for queries.  SAP’s scalable, reliable and secure platform allows you to run your machine learning models in enterprise scenarios as well as serve critical business processes.

Customisable models

Pretrained machine learning (ML) services are a quick way to get started, but sometimes you might need an extra level of adaption. If this is the case, the customisable ML services are just right. You can easily tune existing models with your own training data and create custom models – classifying text contents or images, for example. All this can be done without special data science or ML knowledge.

Image Processing Services

Exploit visual content in pictures and photos. By making images searchable, a computer can automatically detect what is shown in pictures or find similar images.
Previous Next

Customisable image classification

Use a prebuilt image classification model to analyse and tag pictures with generic categories, such as car or food types, or electronic products. Easily create your custom image classification based on your data. This enables the recognition of objects of your choice such as products or spare parts.

Customisable image object detection

Create a custom image object detection to recognise and locate your objects of interest in pictures. It also provides two prebuilt models for detecting faces and humans.

Customisable image feature extraction

Tailor image feature extraction to recognise fine-grained patterns, such as shapes or contours, in images. This service includes prebuilt models for generic images and facial recognition. You can use the results generated for searches or for finding pictures with similar content.

Optical character and scene text recognition

Process document scans and extract text and number data using optical character recognition (OCR). Extract the same information from images with arbitrary fonts, colours, or text orientation using scene text recognition technology.

Image segmentation

Enable the recognition of common consumer product packaging, such as cans or bottles, found in retail stores. This feature provides information such as the location and shape of detected objects and can be used for scenarios such as stock counting or planogram compliance. 

Text Processing Services

Translate and analyse natural language content that is stored in documents, Web sites, or e-mails and reveal its meaning.
Previous Next

Topic detection

Extract topics and keywords from documents and assign the documents to the most relevant topics.

Customisable text and document classification

Use custom text data to create your own text classification service. For example, you can create a service to recognise the sentiment of product reviews or to assess the relevancy of legal documents or research papers.

Text feature extraction

Represent the content of a text document as a vector that can be used for retrieving similar documents.

Language detection

Leverage text recognition technology that detects the language of specific text.


Translate text into different languages. This is especially suited for SAP-related content such as software documentation, manuals or user interfaces to provide accurate terminology.

Speech Processing Services

Establish two-way speech processing capabilities that enable speech-to-text transcription and synthesis.

Previous Next

Speech-to-text services

  • Power voice control for digital assistants or transcribe call centre conversations or phone calls
  • Convert spoken language in videos into text to simplify search and management of contents
  • Recognise more than 100 languages and variants
  • Support real-time recognition of audio streams

Text-to-speech services

  • Promote natural interactions between human users and software
  • Use easy-to-use APIs to enable voice and natural language experiences
  • Pick the most suitable voice from a range of more than 30 voices in 14 languages and variants
  • Generate high-fidelity audio to enable more-natural sounding speech

SAP Software Accessibility

Product Road Map

Back to top