Model comparison
moonshine-base vs whisper-base
These two publish different quantization sets; the table below has the exact sizes.
From the file· summed bytes, KV per layer
Side by side
| moonshine-base | whisper-base | |
|---|---|---|
| Parameters | 62M | 73M |
| Architecture | moonshine | whisper |
| Layers | — | 6 |
| Native context | 194 | — |
| Mixture of experts | no | no |
| Quantizations published | 5 | 19 |
| Smallest quantization | 0.04 GiB | 0.03 GiB |
| Q4_K_M | — | 0.05 GiB |
| Licence | mit | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | moonshine-base | whisper-base | Ratio |
|---|---|---|---|
| 4,096 | — | — | — |
| 8,192 | — | — | — |
| 16,384 | — | — | — |
| 32,768 | — | — | — |
| 65,536 | — | — | — |
| 131,072 | — | — | — |