Curiuous why you deleted all quants just a few hours ago
What was the issue? theyre all gone except 2,3, and 6
What was the issue? theyre all gone except 2,3, and 6
Apologies for the inconvenience. We're in the process of replacing the models with FP16-based version of quantized GGUFs and will re-upload all of them shortly.
great
What was the issue? theyre all gone except 2,3, and 6
Apologies for the inconvenience. We're in the process of replacing the models with FP16-based version of quantized GGUFs and will re-upload all of them shortly.
Ok great! Thanks.. if youve got time for one more question: Im not quite understanding why the it2, t2v, and this version models.. I was under the impression they all use the same workflow?
Thnx for eveytrhting.. this is great
Ok great! Thanks.. if youve got time for one more question: keeping it short: whats the diff with the i2v, t2v, and this verison of models? they all use the same workflow correct?
Here's a quick overview:
T2V (Text-to-Video): Starts from a text prompt to create a video.
📄 Example Workflow
Based on https://huggingface.co/Wan-AI/Wan2.1-T2V-14BI2V (Image-to-Video): Takes one reference image to create a video.
🖼️ Example Workflow
Based on https://huggingface.co/Wan-AI/Wan2.1-I2V-14B-720PPhantom Version: Uses multiple reference images to guide the video, blending different subjects, styles, poses or backgrounds.
🔗 Example Workflow
Based on https://huggingface.co/bytedance-research/Phantom
The workflow setup varies based on the model type.
GGUF models are currently only compatible with the native ComfyUI nodes.
The author of the model also shared a ready-to-use version of each type:
✨ Optimized Workflow which includes both:
- ComfyUI native workflow (GGUF-compatible)
- Wrapper workflow that uses a custom node: ComfyUI-WanVideoWrapper developed by Kijai
Phantom FusionX Q8 - version disappeared!?
Phantom FusionX Q8 - version disappeared!?
Thanks for pointing that out, must have missed it, will re-upload the Q8 version.
EDIT