Instructions to use Julon/Comfy-Pixel-Image-Diffusion-V1.5 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use Julon/Comfy-Pixel-Image-Diffusion-V1.5 with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline from diffusers.utils import load_image # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("Julon/Comfy-Pixel-Image-Diffusion-V1.5", dtype=torch.bfloat16, device_map="cuda") prompt = "Turn this cat into a dog" input_image = load_image("https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/cat.png") image = pipe(image=input_image, prompt=prompt).images[0] - Notebooks
- Google Colab
- Kaggle
PiD v1.5 ComfyUI Finetune
This repository contains the ComfyUI-compatible conversion of the NVIDIA PiD v1.5 undistilled and distilled checkpoints.
Model Type
- Type: Finetune (Undistilled & Distilled)
- Purpose: Native high-resolution decoding for ComfyUI.
Important: ComfyUI Patch
These v1.5 checkpoints require a hidden dimension of 1024. If your ComfyUI implementation is not automatically updated, you must manually patch your installation:
- Navigate to:
ComfyUI/comfy/ldm/pixeldit/pid.py - Set Line 60 and Line 146 to
1024. - Restart ComfyUI.
Conversion
Use the included convert_pid.py script:
python convert_pid.py --input <filename>.pth --output <filename>.safetensors
- Downloads last month
- 33