Quantization publisher

mrfakename

mrfakename publishes 16 quantizations across 1 models in our index, averaging 5.829 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
16
Models covered
1
Avg effective bpw
5.829
across their files

Same model, same quant label, different bytes

largest disagreements first
ModelQuantmrfakenamevsTheirsDifference
mistral-small-3.1-24b-instruct-2503-hfQ2_K8.28 GiBMaziyarPanahi8.28 GiB-0.0%
mistral-small-3.1-24b-instruct-2503-hfQ3_K_L11.55 GiBMaziyarPanahi11.55 GiB-0.0%
mistral-small-3.1-24b-instruct-2503-hfQ3_K_M10.69 GiBMaziyarPanahi10.69 GiB-0.0%
mistral-small-3.1-24b-instruct-2503-hfQ3_K_S9.69 GiBMaziyarPanahi9.69 GiB-0.0%
mistral-small-3.1-24b-instruct-2503-hfQ4_K_S12.62 GiBMaziyarPanahi12.62 GiB-0.0%
mistral-small-3.1-24b-instruct-2503-hfQ5_K_M15.61 GiBMaziyarPanahi15.61 GiB-0.0%
mistral-small-3.1-24b-instruct-2503-hfQ5_K_S15.18 GiBMaziyarPanahi15.18 GiB-0.0%
mistral-small-3.1-24b-instruct-2503-hfQ6_K18.02 GiBMaziyarPanahi18.02 GiB-0.0%
mistral-small-3.1-24b-instruct-2503-hfQ8_023.33 GiBMaziyarPanahi23.33 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
mistral-small-3.1-24b-instruct-2503-hf168.28 GiB