Model comparison

rfdetr vs FireRedPunc

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

From the file· summed bytes, KV per layer

Side by side

rfdetrFireRedPunc
Parameters30M102M
Architecturerfdetrfireredpunc
Layers
Native context
Mixture of expertsnono
Quantizations published123
Smallest quantization0.03 GiB0.05 GiB
Q4_K_M
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
ContextrfdetrFireRedPuncRatio
4,096
8,192
16,384
32,768
65,536
131,072