Model comparison

parakeet-tdt_ctc-110m vs SenseVoiceSmall

At Q4_K_M, parakeet-tdt_ctc-110m is the smaller download — 89,989,600 bytes against 145,738,304.

From the file· summed bytes, KV per layer

Side by side

parakeet-tdt_ctc-110mSenseVoiceSmall
Parameters114M234M
Architectureparakeetsensevoice
Layers
Native context
Mixture of expertsnono
Quantizations published89
Smallest quantization0.07 GiB0.13 GiB
Q4_K_M0.08 GiB0.14 GiB
Licencecc-by-4.0other

KV cache by context

the term that decides long-context viability
Contextparakeet-tdt_ctc-110mSenseVoiceSmallRatio
4,096
8,192
16,384
32,768
65,536
131,072
parakeet-tdt_ctc-110m vs SenseVoiceSmall — size, memory and hardware fit — ossmodeldb