Can I use these with function-based actions?

Yes, you can specify the schema using the @action decorator's 'result' parameter to ensure the function returns the correct data format.

What are Synapse Result Schemas?

They are standardized Python classes used in the Synapse SDK to ensure plugin actions return consistent and validated data structures like training results or inference predictions.

How do I use these schemas in a class-based action?

You assign the desired schema class to the 'result_model' attribute in your action class, which enables automatic validation of the execute method's return value.

What specific result types are supported?

The skill provides schemas for TrainResult, InferenceResult, ExportResult, UploadResult, WeightsResult, and MetricsResult.

Why use schemas instead of plain dictionaries?

Schemas provide type safety, automatic validation, and structured documentation for your plugin's output, making it easier for other components to reliably consume the data.

Synapse Result Schemas

Name: Synapse Result Schemas
Author: datamaker-kr

bydatamaker-kr

•

Data Science & ML

Provides standardized, type-safe result schemas for Synapse SDK plugin actions to ensure consistent data output and validation.

The result-schemas skill streamlines the development of Synapse plugins by providing a set of predefined, validated data models for common AI and data operations. By implementing standardized classes like TrainResult or InferenceResult, developers can ensure their actions return consistent, type-safe data that integrates seamlessly within the Synapse ecosystem. This skill guides users through defining result models in both class-based and function-based actions, significantly reducing errors in return type validation and improving the reliability of machine learning workflows.

Key Features

01Automatic return type validation for Synapse plugin actions

02Support for both class-based and function-based action definitions

03Detailed models for tracking model weights, metrics, and upload statuses

04Simplified integration with Synapse SDK action decorators and base classes

05Standardized result schemas for training, inference, and data export

061 GitHub stars

Use Cases

01Defining the output structure for a machine learning model training action

02Standardizing file upload and data export results across multiple plugin modules

03Validating the response format of an inference service to ensure API consistency

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add datamaker-kr/synapse-claude-marketplace result-schemas

For use in Claude.ai and ChatGPT

Download Skill