Instructions to use LiconStudio/VBVR-wan2.2-comfy-bf16 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use LiconStudio/VBVR-wan2.2-comfy-bf16 with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("LiconStudio/VBVR-wan2.2-comfy-bf16", dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Wan2.2
How to use LiconStudio/VBVR-wan2.2-comfy-bf16 with Wan2.2:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
is the “SNR-Calibrated-Hybrid”I in the document an HiFi FP8 model, and does it need to be used with lightx2Vlora for acceleration?
is the “SNR-Calibrated-Hybrid”I in the document an FP8 model, and does it need to be used with lightx2Vlora for acceleration?
is the “SNR-Calibrated-Hybrid”I in the document an FP8 model, and does it need to be used with lightx2Vlora for acceleration?
The model does not have an acceleration LoRA fused into it. It is the original VBVR-Wan2.2 adapted for ComfyUI. The "SNR-Calibrated-Hybrid" is a version that uses minimal FP8 dynamic quantization based on the Signal-to-Noise Ratio (SNR) to make it compatible with 24GB VRAM.
Thank you for your reply. I have used it with the accelerated LoRA model, and it runs well on the 4060Ti 16G. Thank you for your efforts!!