Downloads · 30 days
94
24% of all-time downloads
scottlowry/Qwen3.5-122B-A10B-oQ3.5e-fp16-mtp
Qwen3.5-122B-A10B-oQ3.5e-fp16-mtp is a machine learning model from scottlowry. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for mlx.
[!NOTE] Updated 2026-11-11: Re-quantized with oMLX v0.7.4.dev1. Please re-download if you have an earlier copy.
Downloads · 30 days
94
24% of all-time downloads
All-time downloads
393
Public
Parameters
125B
64.4 GB on disk
Likes
0
Public
Click a slice to open those files.
.safetensors64.4 GB · 100%
How the weights are stored.
U32125B · 100%
From the Hugging Face model README
[!NOTE] Updated 2026-11-11: Re-quantized with oMLX v0.7.4.dev1. Please re-download if you have an earlier copy.
This model was quantized using oQ (oMLX v0.7.4.dev1) mixed-precision quantization.
The effective average is 4.117 bits/weight for the complete checkpoint.
| Storage format | Logical weights | Share of total | Tensor storage | Effective bits/weight |
|---|---|---|---|---|
| Affine 3-bit, group 64 | 77.309B | 61.805% | 31.500 GiB | 3.50 |
| Affine 4-bit, group 64 | 41.492B | 33.171% | 21.736 GiB | 4.50 |
| Affine 5-bit, group 64 | 1.412B | 1.129% | 0.904 GiB | 5.50 |
| Affine 6-bit, group 64 | 1.219B | 0.974% | 0.922 GiB | 6.50 |
| Affine 8-bit, group 128 | 0.462B | 0.370% | 0.444 GiB | 8.25 |
| Affine 8-bit, group 64 | 2.681B | 2.143% | 2.653 GiB | 8.50 |
| FP16 | 0.060B | 0.048% | 0.111 GiB | 16.00 |
| FP32 | 0.451B | 0.361% | 1.681 GiB | 32.00 |
| Total | 125.086B | 100% | 59.952 GiB | 4.117 |
| Component | Logical weights | Storage |
|---|---|---|
| Language model | 122.112B | 61.115 GB / 56.918 GiB |
| Vision | 0.451B | 1.805 GB / 1.681 GiB |
| MTP head | 2.524B | 1.452 GB / 1.353 GiB |
| Total | 125.086B | 64.373 GB / 59.952 GiB |