Quantization publisher
ChatGpt1
ChatGpt1 publishes 3 quantizations across 1 models in our index, averaging 9.546 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
9.546
across their files
Same model, same quant label, different bytes
largest disagreements first
| Model | Quant | ChatGpt1 | vs | Theirs | Difference |
|---|---|---|---|---|---|
| llama-3-8b-Instruct-bnb-4bit | Q4_K_M | 4.58 GiB | art-from-the-machine | 4.58 GiB | -0.0% |
| llama-3-8b-Instruct-bnb-4bit | F16 | 14.97 GiB | art-from-the-machine | 14.97 GiB | -0.0% |
| llama-3-8b-Instruct-bnb-4bit | Q8_0 | 7.95 GiB | art-from-the-machine | 7.95 GiB | -0.0% |
| llama-3-8b-Instruct-bnb-4bit | Q4_K_M | 4.58 GiB | bartowski | 4.58 GiB | -0.0% |
| llama-3-8b-Instruct-bnb-4bit | Q8_0 | 7.95 GiB | bartowski | 7.95 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
| Model | Quantizations | Smallest |
|---|---|---|
| llama-3-8b-Instruct-bnb-4bit | 3 | 4.58 GiB |