Instructions to use siraxe/H3_Image_experiments with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use siraxe/H3_Image_experiments 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("siraxe/H3_Image_experiments", dtype=torch.bfloat16, device_map="cuda") prompt = "Blend the object into the image seamlessly , (green cyan rimlight:2.7) top lighting,factory building behind , in water , water splashing " 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
import torch
from diffusers import DiffusionPipeline
from diffusers.utils import load_image
# switch to "mps" for apple devices
pipe = DiffusionPipeline.from_pretrained("siraxe/H3_Image_experiments", dtype=torch.bfloat16, device_map="cuda")
prompt = "Blend the object into the image seamlessly , (green cyan rimlight:2.7) top lighting,factory building behind , in water , water splashing "
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]
- Prompt
- Blend the object into the image seamlessly , (green cyan rimlight:2.7) top lighting,factory building behind , in water , water splashing

- Prompt
- Blend the object into the image seamlessly

- Prompt
- Blend the object into the image seamlessly , at night , dark night , (cyan sky:1.4) , top lighting

- Prompt
- Blend the object into the image seamlessly , fix perspective

- Prompt
- change to view selfie shot , keep same identity
H3 Image experiments
1-4.fusion
5.selfie
Fusion loras
v0.1 use at 1.0-1.4
v0.2-v0.3 are stronger , use at 0.4-0.8 for better blending
To use Prompt Weight get https://github.com/kijai/ComfyUI-KJNodes and use MiniMax H3 Prompt Weight node
Trained with https://huggingface.co/ShootTheSound/Fizgig-MiniMax-H3-Image-Training-Adapter
- Downloads last month
- 422