Piccolo-9B-oQ4e

This model was quantized using oQ (oMLX v0.7.4.dev1) mixed-precision quantization.

The effective average is 5.361 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 6.218B 66.081% 3.257 GiB 4.50
Affine 5-bit, group 64 2.625B 27.898% 1.681 GiB 5.50
Affine 6-bit, group 64 0.109B 1.162% 0.083 GiB 6.50
Affine 8-bit, group 64 0.000B 0.003% 0.000 GiB 8.50
BF16 0.457B 4.857% 0.851 GiB 16.00
Total 9.410B 100% 5.873 GiB 5.361

Storage Breakdown

Component Logical weights Storage
Language model 8.954B 5.394 GB / 5.023 GiB
Vision 0.456B 0.912 GB / 0.849 GiB
Total 9.410B 6.306 GB / 5.873 GiB
Downloads last month
63
Safetensors
Model size
9B params
Tensor type
U32
ยท
BF16
ยท
MLX
Hardware compatibility
Log In to add your hardware

4-bit

Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for b2z/Piccolo-9B-oQ4e

Finetuned
Qwen/Qwen3.5-9B
Quantized
(7)
this model