plexe

We have hosted the application plexe in order to run this application in our online workstations with Wine or directly.

Quick description about plexe:

plexe lets you build machine-learning systems from natural-language prompts, turning plain English goals into working pipelines. You describe what you want�a predictor, a classifier, a forecaster�and the tool plans data ingestion, feature preparation, model training, and evaluation automatically. Under the hood an agent executes the plan step by step, surfacing intermediate results and artifacts so you can inspect or override choices. It aims to be production-minded: models can be exported, versioned, and deployed, with reports to explain performance and limitations. The project supports both a Python library and a managed cloud option, meeting teams wherever they prefer to run workloads. The overall goal is to compress the path from idea to usable model while keeping humans in the loop for review and adjustment.

Features:

• Natural-language prompts that generate full ML pipelines end to end
• Agentic execution that handles EDA, feature building, training, and evaluation
• Human-reviewable artifacts, reports, and diffs at each step
• Export, versioning, and deployment paths for production readiness
• Python package for local use plus an optional managed cloud service
• Extensible components so teams can plug in custom data sources or models

Programming Language: Python.
Categories: