Quantization publisher

morikomorizz

morikomorizz publishes 15 quantizations across 2 models in our index, averaging 5.704 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
2
Quantizations
15
Models covered
2
Avg effective bpw
5.704
across their files

Same model, same quant label, different bytes

largest disagreements first
ModelQuantmorikomorizzvsTheirsDifference
GRM-2.6-Plus-0628Q3_K_M15.81 GiBbartowski13.60 GiB+16.2%
GRM-2.6-Plus-0628Q8_028.87 GiBbartowski27.12 GiB+6.5%
GRM-2.6-Plus-0628Q5_K_M20.03 GiBbartowski19.33 GiB+3.6%
GRM-2.6-Plus-0628IQ2_M10.72 GiBbartowski10.13 GiB+5.9%
GRM-2.6-Plus-0628Q4_K_M17.12 GiBbartowski16.55 GiB+3.4%
GRM-2.6-Plus-0628IQ4_NL15.73 GiBbartowski15.20 GiB+3.5%
GRM-2.6-Plus-0628IQ4_XS14.85 GiBbartowski14.50 GiB+2.4%
GRM-2.6-Plus-0628IQ3_M12.78 GiBbartowski12.95 GiB-1.3%
GRM-2.6-Plus-0628BF1650.90 GiBbartowski50.90 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