Ling 3.0 Tiny oQ
Collection
18 items • Updated
Ling-3.0-tiny-oQ4
This model was quantized using oQ (oMLX v0.7.2) mixed-precision quantization.
The effective average is 4.700 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 | 7.433B | 94.170% | 3.894 GiB | 4.50 |
| Affine 5-bit, group 64 | 0.007B | 0.092% | 0.005 GiB | 5.50 |
| Affine 6-bit, group 64 | 0.122B | 1.541% | 0.092 GiB | 6.50 |
| Affine 8-bit, group 64 | 0.331B | 4.190% | 0.327 GiB | 8.50 |
| BF16 | 0.001B | 0.007% | 0.001 GiB | 16.00 |
| FP32 | 0.000B | 0.000% | 0.000 GiB | 32.00 |
| Total | 7.893B | 100% | 4.319 GiB | 4.700 |
Storage Breakdown
| Component | Logical weights | Storage |
|---|---|---|
| Language model | 7.893B | 4.638 GB / 4.319 GiB |
| Total | 7.893B | 4.638 GB / 4.319 GiB |
- Downloads last month
- 18
Model size
8B params
Tensor type
U32
·
BF16 ·
F32 ·
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