Model comparison
Kokoro-82M vs chatterbox
These two publish different quantization sets; the table below has the exact sizes.
From the file· summed bytes, KV per layer
Side by side
| Kokoro-82M | chatterbox | |
|---|---|---|
| Parameters | 82M | 264M |
| Architecture | kokoro | chatterbox-s3gen |
| Layers | 3 | — |
| Native context | — | — |
| Mixture of experts | no | no |
| Quantizations published | 4 | 37 |
| Smallest quantization | 0.13 GiB | 0.17 GiB |
| Q4_K_M | — | 0.46 GiB |
| Licence | apache-2.0 | mit |
KV cache by context
the term that decides long-context viability
| Context | Kokoro-82M | chatterbox | Ratio |
|---|---|---|---|
| 4,096 | — | — | — |
| 8,192 | — | — | — |
| 16,384 | — | — | — |
| 32,768 | — | — | — |
| 65,536 | — | — | — |
| 131,072 | — | — | — |