Model comparison
Wan_2.2_ComfyUI_Repackaged vs HunyuanVideo-I2V
At Q4_K_M, HunyuanVideo-I2V is the smaller download — 7,883,680,512 bytes against 73,954,290,816.
From the file· summed bytes, KV per layer
Side by side
| Wan_2.2_ComfyUI_Repackaged | HunyuanVideo-I2V | |
|---|---|---|
| Parameters | — | 12.8B |
| Architecture | pig | hyvid |
| Layers | — | — |
| Native context | — | — |
| Mixture of experts | no | no |
| Quantizations published | 22 | 24 |
| Smallest quantization | 1.31 GiB | 2.96 GiB |
| Q4_K_M | 68.88 GiB | 7.34 GiB |
| Licence | apache-2.0 | other |
KV cache by context
the term that decides long-context viability
| Context | Wan_2.2_ComfyUI_Repackaged | HunyuanVideo-I2V | Ratio |
|---|---|---|---|
| 4,096 | — | — | — |
| 8,192 | — | — | — |
| 16,384 | — | — | — |
| 32,768 | — | — | — |
| 65,536 | — | — | — |
| 131,072 | — | — | — |