Quantization publisher

darthcrawl

darthcrawl publishes 7 quantizations across 1 models in our index, averaging 5.292 effective bits per weight. Their files differ in size from other publishers' builds of the same nominal quantization on 6 of the pairs we can compare — the same label does not mean the same file.

From the file· summed file bytes
Repositories
1
Quantizations
7
Models covered
1
Avg effective bpw
5.292
across their files

Same model, same quant label, different bytes

largest disagreements first
ModelQuantdarthcrawlvsTheirsDifference
Lilith-31B-v1.0IQ4_XS15.59 GiBmradermacher15.70 GiB-0.7%
Lilith-31B-v1.0Q3_K_M14.24 GiBmradermacher14.24 GiB-0.0%
Lilith-31B-v1.0Q4_K_M17.40 GiBmradermacher17.40 GiB-0.0%
Lilith-31B-v1.0Q5_K_M20.35 GiBmradermacher20.35 GiB-0.0%
Lilith-31B-v1.0Q6_K23.47 GiBmradermacher23.47 GiB-0.0%
Lilith-31B-v1.0Q8_030.39 GiBmradermacher30.39 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
Lilith-31B-v1.0713.43 GiB