Explore huge dataset handling (CV low/high-res and tabular)

[guillaume-byte]

Summary

Explore and validate WeightsLab behavior on huge datasets across computer vision and tabular scenarios.

Scope

• Execute tests on very large datasets for:
  • CV low-resolution
  • CV high-resolution
  • Tabular at scale
• Measure performance, stability, memory pressure, and I/O bottlenecks.
• Capture practical limits and recommended optimizations.

Acceptance Criteria

• Benchmark matrix completed across targeted dataset categories.
• Findings are documented with bottlenecks and observations.
• Actionable recommendations/follow-up tasks are produced.

[guillaume-byte]

For tabular, working with 500k rows, 50cols without frictions