Model comparison
VibeVoice-Realtime-0.5B vs orpheus-3b-0.1-ft
These two publish different quantization sets; the table below has the exact sizes.
From the file· summed bytes, KV per layer
Side by side
| VibeVoice-Realtime-0.5B | orpheus-3b-0.1-ft | |
|---|---|---|
| Parameters | 1.0B | 3.8B |
| Architecture | vibevoice-tts | llama |
| Layers | — | 28 |
| Native context | — | 131,072 |
| Mixture of experts | no | no |
| Quantizations published | 3 | 22 |
| Smallest quantization | 0.65 GiB | 0.91 GiB |
| Q4_K_M | — | 1.95 GiB |
| Licence | mit | apache-2.0 |
KV cache by context
the term that decides long-context viability
| Context | VibeVoice-Realtime-0.5B | orpheus-3b-0.1-ft | Ratio |
|---|---|---|---|
| 4,096 | — | 0.44 GiB | — |
| 8,192 | — | 0.88 GiB | — |
| 16,384 | — | 1.75 GiB | — |
| 32,768 | — | 3.50 GiB | — |
| 65,536 | — | 7.00 GiB | — |
| 131,072 | — | 14.00 GiB | — |