Ornith-1.5-122B oQe MTP
Collection
8 items • Updated
Ornith-1.5-122B-oQ6e-mtp
This model was quantized using oQ (oMLX v0.7.1) mixed-precision quantization.
The effective average is 6.589 bits/weight for the complete checkpoint.
Quantization and Bit Distribution
| Storage format | Logical weights | Share of total | Tensor storage | Effective bits/weight |
|---|---|---|---|---|
| Affine 6-bit, group 128 | 0.906B | 0.724% | 0.659 GiB | 6.25 |
| Affine 6-bit, group 64 | 120.371B | 96.230% | 91.085 GiB | 6.50 |
| Affine 8-bit, group 128 | 0.462B | 0.370% | 0.444 GiB | 8.25 |
| Affine 8-bit, group 64 | 2.836B | 2.267% | 2.806 GiB | 8.50 |
| BF16 | 0.511B | 0.408% | 0.951 GiB | 16.00 |
| Total | 125.086B | 100% | 95.946 GiB | 6.589 |
Storage Breakdown
| Component | Logical weights | Storage |
|---|---|---|
| Language model | 122.112B | 100.043 GB / 93.172 GiB |
| Vision | 0.451B | 0.903 GB / 0.841 GiB |
| MTP head | 2.524B | 2.076 GB / 1.933 GiB |
| Total | 125.086B | 103.021 GB / 95.946 GiB |
- Downloads last month
- 267
Model size
125B params
Tensor type
U32
·
BF16 ·
Hardware compatibility
Log In to add your hardware
6-bit
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Model tree for scottlowry/Ornith-1.5-122B-oQ6e-mtp
Base model
ornith-ai/Ornith-1.5-122B