Quantization publisher

Aqua00

Aqua00 publishes 29 quantizations across 1 models in our index, averaging 4.556 effective bits per weight. Their files differ in size from other publishers' builds of the same nominal quantization on 7 of the pairs we can compare — the same label does not mean the same file.

From the file· summed file bytes
Repositories
1
Quantizations
29
Models covered
1
Avg effective bpw
4.556
across their files

Same model, same quant label, different bytes

largest disagreements first
ModelQuantAqua00vsTheirsDifference
Nemotron-3-Embed-8B-BF16Q3_K_M3.74 GiBAbiray3.74 GiB-0.0%
Nemotron-3-Embed-8B-BF16Q4_K_M4.56 GiBAbiray4.56 GiB-0.0%
Nemotron-3-Embed-8B-BF16Q4_K_S4.33 GiBAbiray4.33 GiB-0.0%
Nemotron-3-Embed-8B-BF16Q5_K_M5.30 GiBAbiray5.30 GiB-0.0%
Nemotron-3-Embed-8B-BF16Q5_K_S5.17 GiBAbiray5.17 GiB-0.0%
Nemotron-3-Embed-8B-BF16Q6_K6.08 GiBAbiray6.08 GiB-0.0%
Nemotron-3-Embed-8B-BF16Q8_07.88 GiBAbiray7.88 GiB-0.0%

A quantization label describes a target, not a recipe. Publishers make different choices about which tensors to keep at higher precision, and some apply an importance matrix while others don't — so two files both honestly labelled the same thing can differ measurably in size and in quality.

Models they publish

ModelQuantizationsSmallest
Nemotron-3-Embed-8B-BF16291.81 GiB