Model comparison

z_image_turbo vs ced-base

These two publish different quantization sets; the table below has the exact sizes.

From the file· summed bytes, KV per layer

Side by side

z_image_turboced-base
Parameters84M86M
Architecturepigced
Layers
Native context
Mixture of expertsnono
Quantizations published173
Smallest quantization0.16 GiB0.12 GiB
Q4_K_M4.20 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
Contextz_image_turboced-baseRatio
4,096
8,192
16,384
32,768
65,536
131,072
z_image_turbo vs ced-base — size, memory and hardware fit — ossmodeldb