Model comparison
parakeet-tdt_ctc-110m vs canary-180m-flash
At Q4_K_M, parakeet-tdt_ctc-110m is the smaller download — 89,989,600 bytes against 139,223,744.
From the file· summed bytes, KV per layer
Side by side
| parakeet-tdt_ctc-110m | canary-180m-flash | |
|---|---|---|
| Parameters | 114M | 189M |
| Architecture | parakeet | canary |
| Layers | — | — |
| Native context | — | — |
| Mixture of experts | no | no |
| Quantizations published | 8 | 6 |
| Smallest quantization | 0.07 GiB | 0.13 GiB |
| Q4_K_M | 0.08 GiB | 0.13 GiB |
| Licence | cc-by-4.0 | cc-by-4.0 |
KV cache by context
the term that decides long-context viability
| Context | parakeet-tdt_ctc-110m | canary-180m-flash | Ratio |
|---|---|---|---|
| 4,096 | — | — | — |
| 8,192 | — | — | — |
| 16,384 | — | — | — |
| 32,768 | — | — | — |
| 65,536 | — | — | — |
| 131,072 | — | — | — |