Model comparison
moonshine-streaming-medium vs canary-180m-flash
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-streaming-medium | canary-180m-flash | |
|---|---|---|
| Parameters | 266M | 189M |
| Architecture | moonshine_streaming | canary |
| Layers | 14 | — |
| Native context | 4,096 | — |
| Mixture of experts | no | no |
| Quantizations published | 3 | 6 |
| Smallest quantization | 0.28 GiB | 0.13 GiB |
| Q4_K_M | — | 0.13 GiB |
| Licence | mit | cc-by-4.0 |
KV cache by context
the term that decides long-context viability
| Context | moonshine-streaming-medium | canary-180m-flash | Ratio |
|---|---|---|---|
| 4,096 | 0.14 GiB | — | — |
| 8,192 | 0.27 GiB | — | — |
| 16,384 | 0.55 GiB | — | — |
| 32,768 | 1.09 GiB | — | — |
| 65,536 | 2.19 GiB | — | — |
| 131,072 | 4.38 GiB | — | — |