embeddinggemma-2-oQ5e

This model was quantized using oQ (oMLX v0.7.1) mixed-precision quantization.

The effective average is 12.356 bits/weight for the complete checkpoint.

Quantization and Bit Distribution

Storage format	Logical weights	Share of total	Tensor storage	Effective bits/weight
Affine 5-bit, group 64	0.203B	27.223%	0.130 GiB	5.50
Affine 6-bit, group 64	0.032B	4.261%	0.024 GiB	6.50
Affine 8-bit, group 64	0.038B	5.071%	0.037 GiB	8.50
BF16	0.472B	63.445%	0.880 GiB	16.00
Total	0.744B	100%	1.071 GiB	12.356

Storage Breakdown

Component	Logical weights	Storage
Language model	0.272B	0.205 GB / 0.191 GiB
Vision	0.167B	0.335 GB / 0.312 GiB
Audio	0.305B	0.610 GB / 0.568 GiB
Total	0.744B	1.150 GB / 1.071 GiB