Model comparison

Anima vs stable-diffusion-v1-5

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

From the file· summed bytes, KV per layer

Side by side

Animastable-diffusion-v1-5
Parameters2.1B860M
Architecturecosmos
Layers
Native context
Mixture of expertsnono
Quantizations published237
Smallest quantization0.95 GiB1.46 GiB
Q4_K_M1.29 GiB
Licenceothercreativeml-openrail-m

KV cache by context

the term that decides long-context viability
ContextAnimastable-diffusion-v1-5Ratio
4,096
8,192
16,384
32,768
65,536
131,072