Instructions to use lilylilith/QI_2.1_AnyAngle with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use lilylilith/QI_2.1_AnyAngle 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("Qwen/Qwen-Image-2.1", dtype=torch.bfloat16, device_map="cuda") pipe.load_lora_weights("lilylilith/QI_2.1_AnyAngle") 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
- Local Apps Settings
- Draw Things
Add a fully native ComfyUI workflow for AnyAngle (Gaussian Splat + camera render inside ComfyUI, no Blender)
#2
by ComfyUI-Wiki - opened
What this PR adds
A ready-to-run, fully native ComfyUI workflow for the QI2.1 AnyAngle LoRA: the whole AnyAngle process runs inside ComfyUI, with no extra dependencies and no custom nodes, so nothing has to be exported and there is no need to switch to Blender or any other software.
Files:
workflows/image_qwen_image_2_1_any_angle.jsonβ the workflow (drag it into ComfyUI)workflows/workflow_screenshot.pngβ node graph overviewworkflows/output_example.webpβ example: input | coarse splat render | final outputREADME.mdβ new "Fully Native ComfyUI Workflow (No Blender Required)" section
Requires a current ComfyUI: the splat / camera nodes and the Qwen Image 2.1 nodes are part of current core, and the workflow uses subgraphs (recent frontend). The model files and their target folders are listed in the note inside the workflow itself.
Only README.md is edited (one new section inserted after Usage) and the new workflows/ folder is added β no existing content was removed or reworded.