Hunyuan I2V Native Workflow is Tied to "Unfixed" Model

### Expected Behavior

I posted this on Kijai's repo as well (https://github.com/kijai/ComfyUI-HunyuanVideoWrapper/issues/462), but it looks like the workflows and tutorials for the native implementation of Hunyuan I2V is still set up around the first-day bugged release of the model, which ends up looking closer to T2V (loss of initial frame). Switching the model in the workflow to the "replaced"/"fixed" one isn't enough. From messing around, it seems like it gets closer to the i2v repo output with changing guidance type to v2 replace, but I'm not sure if that's what we're supposed to do here since I'm not familiar with what that does.

### Actual Behavior

Video output with Hunyuan I2V replaced model looks like T2V (loses initial frame reference), similar to the unfixed model behavior.

### Steps to Reproduce

Download native workflow from https://comfyui-wiki.com/en/tutorial/advanced/hunyuan-image-to-video-workflow-guide-and-example#google_vignette
Change model to "replaced" model from huggingface

### Debug Logs

```powershell
n/a
```

### Other

_No response_