Quantization publisher

dranger003

dranger003 publishes 47 quantizations across 2 models in our index, averaging 4.463 effective bits per weight. Their files differ in size from other publishers' builds of the same nominal quantization on 8 of the pairs we can compare — the same label does not mean the same file.

From the file· summed file bytes
Repositories
2
Quantizations
47
Models covered
2
Avg effective bpw
4.463
across their files

Same model, same quant label, different bytes

largest disagreements first
ModelQuantdranger003vsTheirsDifference
c4ai-command-r-v01Q2_K12.87 GiBDavidAU12.87 GiB0.0%
c4ai-command-r-plusQ3_K_M47.48 GiBpmysl47.48 GiB0.0%
c4ai-command-r-plusQ5_K_M68.57 GiBpmysl68.57 GiB0.0%
c4ai-command-r-plusQ5_K_S66.87 GiBpmysl66.87 GiB0.0%
c4ai-command-r-plusQ3_K_L51.60 GiBpmysl51.60 GiB0.0%
c4ai-command-r-plusQ2_K36.78 GiBpmysl36.78 GiB0.0%
c4ai-command-r-plusQ4_K_M58.44 GiBpmysl58.44 GiB0.0%
c4ai-command-r-plusQ4_K_S55.55 GiBpmysl55.55 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
c4ai-command-r-v01247.94 GiB
c4ai-command-r-plus2321.59 GiB