clef-oQ4e
This model was quantized using oQ (oMLX v0.7.1rc1) mixed-precision quantization.
The effective average is 4.936 bits/weight for the complete checkpoint.
Quantization and Bit Distribution
| Storage format | Logical weights | Share of total | Tensor storage | Effective bits/weight |
|---|---|---|---|---|
| Affine 4-bit, group 64 | 22.986B | 83.633% | 12.042 GiB | 4.50 |
| Affine 5-bit, group 64 | 2.636B | 9.589% | 1.687 GiB | 5.50 |
| Affine 6-bit, group 64 | 1.271B | 4.626% | 0.962 GiB | 6.50 |
| BF16 | 0.591B | 2.152% | 1.102 GiB | 16.00 |
| Total | 27.485B | 100% | 15.793 GiB | 4.936 |
Storage Breakdown
| Component | Logical weights | Storage |
|---|---|---|
| Language model | 26.896B | 15.780 GB / 14.696 GiB |
| Vision | 0.461B | 0.921 GB / 0.858 GiB |
| Joint schema head | 0.128B | 0.256 GB / 0.239 GiB |
| Total | 27.485B | 16.958 GB / 15.793 GiB |
- Downloads last month
- 18
Model size
27B params
Tensor type
U32
·
BF16 ·
Hardware compatibility
Log In to add your hardware
4-bit
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support