Model comparison

orpheus-3b-0.1-pretrained vs orpheus-3b-0.1-ft

At Q4_K_M, orpheus-3b-0.1-ft is the smaller download — 2,092,567,712 bytes against 2,363,760,640.

From the file· summed bytes, KV per layer

Side by side

orpheus-3b-0.1-pretrainedorpheus-3b-0.1-ft
Parameters3.8B3.8B
Architecturellamallama
Layers2828
Native context131,072131,072
Mixture of expertsnono
Quantizations published4522
Smallest quantization1.40 GiB0.91 GiB
Q4_K_M2.20 GiB1.95 GiB
Licenceapache-2.0apache-2.0

KV cache by context

the term that decides long-context viability
Contextorpheus-3b-0.1-pretrainedorpheus-3b-0.1-ftRatio
4,0960.44 GiB0.44 GiB1.00×
8,1920.88 GiB0.88 GiB1.00×
16,3841.75 GiB1.75 GiB1.00×
32,7683.50 GiB3.50 GiB1.00×
65,5367.00 GiB7.00 GiB1.00×
131,07214.00 GiB14.00 GiB1.00×