Quantization publisher

art-from-the-machine

art-from-the-machine publishes 3 quantizations across 1 models in our index, averaging 9.547 effective bits per weight. Their files differ in size from other publishers' builds of the same nominal quantization on 5 of the pairs we can compare — the same label does not mean the same file.

From the file· summed file bytes
Repositories
1
Quantizations
3
Models covered
1
Avg effective bpw
9.547
across their files

Same model, same quant label, different bytes

largest disagreements first
ModelQuantart-from-the-machinevsTheirsDifference
llama-3-8b-Instruct-bnb-4bitQ4_K_M4.58 GiBChatGpt14.58 GiB0.0%
llama-3-8b-Instruct-bnb-4bitF1614.97 GiBChatGpt114.97 GiB0.0%
llama-3-8b-Instruct-bnb-4bitQ8_07.95 GiBChatGpt17.95 GiB0.0%
llama-3-8b-Instruct-bnb-4bitQ4_K_M4.58 GiBbartowski4.58 GiB0.0%
llama-3-8b-Instruct-bnb-4bitQ8_07.95 GiBbartowski7.95 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
llama-3-8b-Instruct-bnb-4bit34.58 GiB