Quantization publisher

Vikhrmodels

Vikhrmodels publishes 33 quantizations across 1 models in our index, averaging 6.156 effective bits per weight. Their files differ in size from other publishers' builds of the same nominal quantization on 9 of the pairs we can compare — the same label does not mean the same file.

From the file· summed file bytes
Repositories
1
Quantizations
33
Models covered
1
Avg effective bpw
6.156
across their files

Same model, same quant label, different bytes

largest disagreements first
ModelQuantVikhrmodelsvsTheirsDifference
Vikhr-Gemma-2B-instructQ8_02.59 GiBbartowski2.59 GiB-0.0%
Vikhr-Gemma-2B-instructQ5_K_M1.79 GiBbartowski1.79 GiB-0.0%
Vikhr-Gemma-2B-instructQ5_K_S1.75 GiBbartowski1.75 GiB-0.0%
Vikhr-Gemma-2B-instructQ6_K2.00 GiBbartowski2.00 GiB-0.0%
Vikhr-Gemma-2B-instructQ4_K_M1.59 GiBbartowski1.59 GiB-0.0%
Vikhr-Gemma-2B-instructQ4_K_S1.53 GiBbartowski1.53 GiB-0.0%
Vikhr-Gemma-2B-instructIQ4_XS1.46 GiBbartowski1.46 GiB0.0%
Vikhr-Gemma-2B-instructQ3_K_L1.44 GiBbartowski1.44 GiB0.0%
Vikhr-Gemma-2B-instructIQ3_M1.30 GiBbartowski1.30 GiB0.0%

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
Vikhr-Gemma-2B-instruct330.78 GiB