Model comparison

ERNIE-Image vs FLUX.2-klein-9B

At Q4_K_M, ERNIE-Image is the smaller download — 2,146,489,664 bytes against 5,027,783,648.

From the file· summed bytes, KV per layer

Side by side

ERNIE-ImageFLUX.2-klein-9B
Parameters8.0B9.1B
Architecturewanflux
Layers
Native context
Mixture of expertsnono
Quantizations published2929
Smallest quantization2.00 GiB3.71 GiB
Q4_K_M2.00 GiB4.68 GiB
Licenceapache-2.0other

KV cache by context

the term that decides long-context viability
ContextERNIE-ImageFLUX.2-klein-9BRatio
4,096
8,192
16,384
32,768
65,536
131,072