Model comparison

FireRedPunc vs TinyMistral-248M-v2-Instruct

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

From the file· summed bytes, KV per layer

Side by side

FireRedPuncTinyMistral-248M-v2-Instruct
Parameters102M248M
Architecturefireredpuncllama
Layers12
Native context32,768
Mixture of expertsnono
Quantizations published36
Smallest quantization0.05 GiB0.10 GiB
Q4_K_M0.14 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
ContextFireRedPuncTinyMistral-248M-v2-InstructRatio
4,096
8,192
16,384
32,768
65,536
131,072