RunForgeDesktop.Core 1.0.2

<p align="center"> <a href="README.ja.md">日本語</a> | <a href="README.zh.md">中文</a> | <a href="README.es.md">Español</a> | <a href="README.fr.md">Français</a> | <a href="README.hi.md">हिन्दी</a> | <a href="README.it.md">Italiano</a> | <a href="README.pt-BR.md">Português (BR)</a> </p>

<p align="center"><img src="https://raw.githubusercontent.com/mcp-tool-shop-org/brand/main/logos/runforge-desktop/readme.png" alt="RunForge Desktop" width="400"></p>

<p align="center"> <a href="https://www.nuget.org/packages/RunForgeDesktop.Core"><img src="https://img.shields.io/nuget/v/RunForgeDesktop.Core?label=RunForgeDesktop.Core" alt="NuGet"></a> <a href="LICENSE"><img src="https://img.shields.io/badge/License-MIT-blue.svg" alt="License: MIT"></a> <a href="https://github.com/mcp-tool-shop-org/runforge-desktop/releases"><img src="https://img.shields.io/badge/platform-Windows%2010%2F11-0078D6?logo=windows" alt="Platform"></a> <a href="https://mcp-tool-shop-org.github.io/runforge-desktop/"><img src="https://img.shields.io/badge/Landing_Page-live-blue" alt="Landing Page"></a> </p>

RunForge Desktop is a Windows-native desktop application for creating, monitoring, and inspecting ML training runs.

It provides a visual control plane for ML experiments—creating runs, monitoring live training progress with real-time charts, and browsing completed runs with full artifact inspection.

Canonical upstream (artifacts, schemas, guarantees): https://github.com/mcp-tool-shop-org/runforge-vscode

Why

Most ML experiment trackers are cloud-first, SaaS platforms that require accounts, send telemetry, and add complexity. RunForge Desktop takes the opposite approach: everything runs locally on your machine.

With RunForge Desktop you can:

• Create training runs with preset configurations
• Monitor live training with real-time charts and logs
• Browse completed runs and their outputs
• Inspect metrics, logs, and artifacts
• Manage runs (cancel, view outputs, copy commands)

All training runs locally on your machine using Python. No cloud. No telemetry. No accounts.

NuGet Packages

dotnet add package RunForgeDesktop.Core

Quick Start

Installation

Option 1: MSIX Package (Recommended)

1. Download the .msix file from Releases
2. Double-click to install
3. Launch from Start Menu

Option 2: Build from Source

git clone https://github.com/mcp-tool-shop-org/runforge-desktop
cd runforge-desktop
dotnet run --project src/RunForgeDesktop/RunForgeDesktop.csproj

See docs/INSTALL.md for detailed installation options.

Usage

1. Launch RunForge Desktop
2. Select Workspace - Click "Select Workspace" and choose a folder for your ML experiments
3. Start Training - Click "Train" to configure and launch a training run
4. Monitor Live - Watch training progress with real-time loss charts and logs
5. Browse Runs - View all runs with filtering by status
6. Inspect Details - Click any run to view metrics, artifacts, and outputs

Features

Training Run Creation

• Configure training runs with epoch presets (Quick, Standard, Extended, Custom)
• GPU/CPU device selection with automatic detection
• Advanced settings: batch size, learning rate, optimizer, scheduler
• Optional custom dataset path

Hyperparameter Sweeps (MultiRun)

• Run multiple experiments with different hyperparameter combinations
• Configure learning rates, batch sizes, and optimizers as comma-separated lists
• Automatic grid search across all combinations
• Track best-performing configuration by final loss

Live Monitoring

• Real-time loss chart with automatic updates
• Live log streaming from training process
• Progress tracking (epoch, step, elapsed time)
• Cancel running training at any time

Run Browsing

• Browse runs with newest-first ordering
• Filter by status: Pending, Running, Completed, Failed, Cancelled
• View run details and outputs

Run Inspection

• Metrics - Loss curves, accuracy, training statistics
• Logs - Full stdout/stderr from training process
• Artifacts - Open output folder, copy training command

Diagnostics

• View app version, framework, and memory usage
• View workspace path and Python configuration
• Copy diagnostics to clipboard for support

Core Principles

Local-first

All training runs on your machine. No cloud required.

Transparent

See exactly what's happening: live logs, real-time metrics, full process control.

Simple

One workspace, clear presets, no configuration files to manage.

Auditable

All run artifacts saved to disk for inspection and reproducibility.

How It Works

RunForge Desktop
  │
  ├── Select Workspace (any folder)
  │
  ├── Create Run (preset + device + optional dataset)
  │
  ├── Spawn Python training process
  │
  ▼
.ml/
  └── runs/
      └── 20240101-123456-myrun-abc1/
          ├── run.json       (manifest)
          ├── metrics.jsonl  (live metrics)
          ├── stdout.log     (live logs)
          └── stderr.log     (errors)

RunForge Desktop manages the full lifecycle: creation, execution, monitoring, and inspection.

System Requirements

Platform & Packaging

The app follows standard Windows permission models for file access.

Project Status

See RELEASE_NOTES_v0.4.0.md for recent changes.

Development

Prerequisites

• .NET 10 SDK
• Windows 10/11
• Visual Studio 2022 (17.12+) with MAUI workload, OR VS Code with .NET MAUI extension

Build

# Debug build
dotnet build

# Run tests
dotnet test

# Release build
.\scripts\build-release.cmd

Project Structure

runforge-desktop/
├── src/
│   ├── RunForgeDesktop/          # MAUI app (UI, ViewModels)
│   └── RunForgeDesktop.Core/     # Core services, models
├── tests/
│   └── RunForgeDesktop.Core.Tests/
├── docs/
│   ├── PHASE-DESKTOP-0.1-ACCEPTANCE.md
│   └── INSTALL.md
└── scripts/
    ├── build-msix.ps1
    └── build-release.cmd

Relationship to RunForge Core

All schemas, guarantees, and artifact formats are defined and frozen in:

https://github.com/mcp-tool-shop-org/runforge-vscode

This repository contains:

• No training logic
• No schema definitions
• No contract ownership

RunForge Desktop consumes those artifacts faithfully.

Intended Audience

• Developers training models locally on Windows
• Researchers who need simple, inspectable experiment tracking
• Anyone who wants a native Windows ML training UI
• Teams that want local-first, no-cloud ML workflows

License

MIT License - See LICENSE for details.

Reliability Gauntlets

RunForge ships with a repeatable reliability suite you can run locally to validate queueing, pause/resume, cancellation, crash recovery, fairness, disk drift resilience, and Desktop reconnect behavior.

See: docs/GAUNTLETS.md

Contributing

Contributions welcome. Please respect the core principles:

• Keep it simple and local-first
• No cloud dependencies or telemetry
• Clear, actionable error messages

Support

• Issues: GitHub Issues
• Diagnostics: Use the Diagnostics page to copy system info for bug reports