Sweet Software Practices Launches Free Windows Application for ONNX Batch Image Classification

eTradeWire News/10851891
Dulce™ enables local batch image classification using existing ONNX models, without requiring custom software, cloud or MLOps platform.

OSHKOSH, Wis. - eTradeWire -- Sweet Software Practices announces the availability of Dulce™ Batch Image Classification for Windows®, a desktop application that enables users to run their own ONNX image-classification models against folders of images directly on their Windows PCs.

Available through the Microsoft Store, Dulce Free is permanently free for personal and commercial use.

Dulce addresses a practical challenge for users working with trained image-classification models: executing those models against existing image collections without developing, testing, deploying, and maintaining a separate inference application.

The software provides a graphical interface for configuring compatible ONNX models, defining image preprocessing parameters, selecting image folders, and executing batch classification jobs. Results include per-image predictions, structured output files, execution telemetry, and run summaries.

All processing occurs locally. Images and models are not uploaded to a cloud inference service, and users do not need to deploy Python environments, containers, inference or MLOps platforms.

More on eTradeWire News
Designed for Model Developers and Technical Users

Dulce serves several potential applications, including:
  • ONNX model developers who need a Windows application through which end users can execute their classifiers.
  • Machine- and computer-vision engineers processing existing image datasets.
  • Manufacturing quality and R&D teams conducting offline image-classification studies.
  • Organizations reprocessing archived inspection images or evaluating updated classification models.
Because Dulce accepts compatible, separately supplied ONNX models, it can support different classification applications without requiring a custom inference application for each model.

Engineered for Reliable Local Batch Processing

Unlike a one-off inference script, Dulce is a dedicated Windows application backed by a comprehensive suite of 2,577 unit tests. Its integrated model configuration, preprocessing, execution, and reporting capabilities reduce the need for developers to build, test, and maintain custom batch-inference software.

Rather than requiring users to submit images to a centralized inference server, Dulce executes classification jobs directly on their Windows PCs. The Standard edition can distribute batch workloads concurrently across selected, supported CPU, DirectML, and NVIDIA CUDA devices, allowing users to take advantage of existing local computing hardware without deploying a shared inference service.

More on eTradeWire News
This approach allows model developers and engineering teams to concentrate on developing and validating their classification models rather than building and maintaining the software infrastructure needed to execute them.

Free and Standard Editions

Dulce Free
includes CPU-based inference, model profiles, batch execution, classification results, and reporting. It is permanently free for both commercial and personal use.

For users requiring GPU execution, Dulce Standard is available as an optional $149.99 one-time upgrade. Standard adds DirectML and NVIDIA CUDA support, including concurrent execution across supported CPU and GPU devices.

Availability

Dulce Free is available now through the Microsoft Store.

https://dulcebatch.com/

Contact
Michael Frank
Sweet Software Practices Corporation
***@sweetsoftwarepractices.com


Source: Sweet Software Practices Corporation

Show All News | Disclaimer | Report Violation

0 Comments

Latest on eTradeWire News