Gemma 4 12B exposure continuation training — peak dedicated GPU

Gemma 4 12B exposure continuation training — peak dedicated GPU: 9.15 GiB on a 12 GB card.

Single reading

9,824,735,232 bytes

MEASURED recorded 22 September 2026

Protocol

Test definition

GPU memory telemetry from the project's watcher / allocator counters. 547 samples; shared 85,983,232; no host fallback; exit 0.

Limits

Limitations

  • Sampled gauges; sub-sample transients not exhaustively proven. Physical card is 12,227 MiB; allocator and dedicated readings are different counters.

Log

Recorded values

Date Value Evidence Samples Source
22 September 2026 9,824,735,232 bytes MEASURED lm-runtime/reports/production-architecture-study-001-gemma4-bounded-adapter-results-006/bounded-adapter-watch-receipt.json

Source paths refer to the private JARVIS repository and runtime records. They are listed for audit and are not published.