Quantization publisher

vcruz305

vcruz305 publishes 39 quantizations across 5 models in our index, averaging 5.841 effective bits per weight. Their files differ in size from other publishers' builds of the same nominal quantization on 15 of the pairs we can compare — the same label does not mean the same file.

From the file· summed file bytes
Repositories
5
Quantizations
39
Models covered
5
Avg effective bpw
5.841
across their files

Same model, same quant label, different bytes

largest disagreements first
ModelQuantvcruz305vsTheirsDifference
Hy3IQ1_M85.45 GiBbartowski66.22 GiB+29.1%
Hy3IQ3_XXS109.28 GiBbartowski117.43 GiB-6.9%
Hy3Q3_K_M132.94 GiBAngelSlim125.35 GiB+6.0%
Hy3Q3_K_M132.94 GiBbartowski128.10 GiB+3.8%
Solar-Open2-250BQ2_K85.18 GiBprometheusAIR88.87 GiB-4.2%
Solar-Open2-250BIQ1_M52.67 GiBprometheusAIR56.30 GiB-6.4%
Solar-Open2-250BIQ4_XS124.46 GiBprometheusAIR126.88 GiB-1.9%
Hy3IQ1_M85.45 GiBAngelSlim83.30 GiB+2.6%
Hy3IQ2_M93.14 GiBbartowski95.08 GiB-2.0%
Hy3Q2_K101.28 GiBbartowski99.36 GiB+1.9%
Hy3IQ4_XS148.22 GiBbartowski149.73 GiB-1.0%
Hy3Q4_K_M168.57 GiBAngelSlim169.81 GiB-0.7%
Hy3Q4_K_M168.57 GiBbartowski169.65 GiB-0.6%
Hy3Q5_K_M197.61 GiBbartowski198.22 GiB-0.3%
Hy3Q8_0295.84 GiBbartowski295.84 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
Hy3985.45 GiB
Ornith-1.0-35B-AEON-Ultimate-Uncensored-BF1697.67 GiB
Bonsai-27B-unpacked612.39 GiB
Solar-Open2-250B952.67 GiB
Ternary-Bonsai-27B-unpacked612.39 GiB