webxos/underworld_dataset_v3

   

<div style=" background: #00FF00; border-left: 4px solid #00FF00; padding: 1.5rem; margin: 2rem 0; font-family: 'Fira Code', 'Courier New', monospace; color: #00FF00; border-radius: 0 8px 8px 0; "> <pre style=" font-size: 12px; line-height: 1.2; margin: 0; overflow-x: auto; color: #00FF00; "> _____ _________ _________ _____ | | | | \ | | \ __| _ \ | | || | __ \ | | \ | | | | \| | | | | |_ | |/ / | | || | | | |/ / | | | | | | | | | . ` | | | | || /| |/\| || | | | /| | | | | | | || | |\ | |/ /| |__| |\ \\ /\ /\ \/ / |\ \| |__| |/ / \_/\| \/_/ \__/\| \|\/ \/ \_/\| \\___/__/

</div>

UNDERWORLD Dataset v3

• —Visualizes the Fast Inverse Square Root (FISR / Quake III) algorithm.
• —120 rows total (train split only).
• —Main content: 1280px PNG image frames showing bit hacks, Newton-Raphson steps, error surfaces, 3D math plots.
• —Numerical_data.csv (regression), metadata.json (conditional).
• —Size ~3.5 MB. Generated via UNDERGROUND: FISR tool (downloadable in repo).
• —Magic Number: 0x5f23aac5
• —Newton Iterations: 3
• —Input Range: 0.1 to 1000
• —Maximum Error: 1.1742636926798086e+287%

Generated with UNDERWORLD: FISR by webXOS (credits: John Carmack for FISR), Educational visualization of the Quake III Arena optimization algorithm.

The UNDERWORLD app by webXOS is available for download in the /underworld/ folder of this repo so users can create their own datasets.

Use cases:

• —Training ML models for fault detection / anomaly detection in time-series or sensor data.

• —Simulating hardware faults (bit flips, stuck-at, etc.) for robust AI / embedded ML.

• —Reliability engineering: predict system failures under errors.

• —Synthetic data for safety-critical systems (automotive, aerospace, IoT) where real fault data is rare.

• —Benchmarking error-correction / resilient algorithms.

• —Visual sequence learning → train models on math visualization sequences (frame prediction, video understanding).

• —Image-to-text / captioning → describe FISR steps from images.

• —Visual question answering → QA on algorithm visuals.

• —Regression from images → predict error metrics from visualization frames.

• —Educational multimodal models → teach bit manipulation / fast math approx.

• —Conditional generation → use metadata to condition on input range/error.

• —3D math function visualization benchmark → compare rendering / understanding.

Educational Purposes:

1. 1.The Fast Inverse Square Root algorithm implementation
2. 2.Error analysis of the approximation
3. 3.3D visualization of mathematical functions
4. 4.Bit-level manipulation techniques

For Training:

1. 1.Use frames/ for visual sequence learning
2. 2.Use numerical_data.csv for regression tasks
3. 3.Use metadata.json for conditional generation
4. 4.Train models to understand optimization algorithms

License:

MIT