Do I need to train a machine learning model to use this skill?

No, SAP-RPT-1-OSS uses in-context learning, which allows it to make predictions on new data based on provided examples without a separate training or fine-tuning phase.

Can this skill handle unstructured data like images or PDFs?

No, this skill is specifically designed for structured tabular data, such as CSV exports or DataFrames from SAP systems.

What are the hardware requirements for running this skill?

While it can run on a CPU for testing, a GPU with at least 24GB VRAM (like an RTX 4090) is recommended for development, and 80GB (like an A100) is optimal for production accuracy.

How does the model understand SAP technical field names?

The skill includes data preparation scripts that map technical SAP codes (e.g., BUKRS) to semantic business terms (e.g., COMPANY_CODE) to help the model's LLM-based embedding layer understand the context.

Is my data sent to a third-party cloud service?

The SAP-RPT-1-OSS model runs locally on your infrastructure via Hugging Face. However, there is an optional API client included if you prefer to use the SAP RPT Playground service.

SAP RPT-1 Predictor

Name: SAP RPT-1 Predictor
Author: amitlals

byamitlals

•

データサイエンスとML

Leverages SAP's open-source tabular foundation model to perform predictive analytics on structured business data without model training.

The SAP RPT-1 Predictor skill empowers Claude to execute sophisticated classification and regression tasks on SAP business data exports by utilizing SAP-RPT-1-OSS. This tabular foundation model uses in-context learning to predict field values directly from table rows, eliminating the need for traditional model training. It is specifically optimized for business use cases such as forecasting customer churn, predicting delivery delays, and detecting financial anomalies. By integrating semantic mapping for SAP technical fields, it ensures high accuracy and provides a seamless workflow for turning SAP exports into actionable predictive insights.

主な機能

01Scalable performance with support for bagging and large context sizes

02Automated semantic renaming for SAP technical field names

03Local execution via Hugging Face for data privacy and security

04Support for classification and regression tasks on tabular data

05In-context learning for zero-shot and few-shot predictive analytics

062 GitHub stars

ユースケース

01Identifying anomalies and fraud in financial journal entries (ACDOCA)

02Predicting customer churn and payment default risks in FI/SD modules

03Forecasting delivery delays and production risks for supply chain management

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add amitlals/sap-rpt1-oss-predictor sap-rpt1-oss-predictor

For use in Claude.ai and ChatGPT

Download Skill