Quantization publisher
Xviers
Xviers publishes 13 quantizations across 1 models in our index, averaging 5.549 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
13
Models covered
1
Avg effective bpw
5.549
across their files
Same model, same quant label, different bytes
largest disagreements first
| Model | Quant | Xviers | vs | Theirs | Difference |
|---|---|---|---|---|---|
| whisper-large-v3-turbo | Q8_0 | 0.81 GiB | handy-computer | 0.83 GiB | -2.5% |
| whisper-large-v3-turbo | Q6_K | 0.63 GiB | handy-computer | 0.64 GiB | -1.8% |
| whisper-large-v3-turbo | Q8_0 | 0.81 GiB | handy-computer | 0.83 GiB | -1.4% |
| whisper-large-v3-turbo | Q4_0 | 0.44 GiB | oxide-lab | 0.43 GiB | +2.4% |
| whisper-large-v3-turbo | Q4_0 | 0.43 GiB | oxide-lab | 0.44 GiB | -2.3% |
| whisper-large-v3-turbo | Q4_1 | 0.49 GiB | oxide-lab | 0.48 GiB | +2.1% |
| whisper-large-v3-turbo | Q4_1 | 0.48 GiB | oxide-lab | 0.49 GiB | -2.1% |
| whisper-large-v3-turbo | Q8_0 | 0.81 GiB | oxide-lab | 0.81 GiB | -1.1% |
| whisper-large-v3-turbo | Q8_0 | 0.81 GiB | oxide-lab | 0.81 GiB | +1.1% |
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 |
|---|---|---|
| whisper-large-v3-turbo | 13 | 0.27 GiB |