Weight Quantization Footprints

Parent: 16 GB VRAM residency budgeting · Published reference · snapshot 2026-09-30

Source-anchored research for Weight Quantization Footprints. Run status: COMPLETED.

These notes link each claim to its source. Some citations identify archived research reports that are no longer hosted on this site. A published reference means the content is available; it does not certify independent review or accuracy.Read the editorial policy and follow the sources before relying on a claim.

Measurements and reference values

A 14-billion parameter model: FP16 - 28.0 GB, FP8/INT8 - 14.0 GB, 4-bit (Q4_K_M) - 7.62 GB (high confidence) [source]
On 16 GB RTX 5080: Q4_K_M leaves 8.38 GB for KV cache and scratch (high confidence) [source]

Children

No children recorded.

← the whole tree · 3D view· how to read this page