Yuanshi9815
commented
20 hours ago Thank you so much for sharing these detailed instructions and improvements! 🙏
The grid coordinates feature is a clever addition that can help users with precise masking. We've pinned your usage guide so other community members can easily find and reference these helpful enhancements when working with the inpainting notebook.
We appreciate community members like you who take the time to share their experiences with the project! 💫
if you use inpainting.ipynb with jupyter notebook 1.) i did upgrade to the file. and now you have grid with numcer of the coordinates 2.) add save output picture to folder
instructions
after this cell:
pipe.load_lora_weights( "Yuanshi/OminiControl", weight_name=f"experimental/fill.safetensors", adapter_name="fill", )
make new cell and put:
image = Image.open(r"H:\GITHUB\OMINI CONTROL\assets\monalisa.jpg").convert("RGB").resize((512, 512)) from matplotlib import pyplot as plt plt.imshow(image) plt.xticks(range(0, 512, 32)) # חלוקה כל 32 פיקסלים plt.yticks(range(0, 512, 32)) plt.grid(True) plt.show()
and after that make a new cell and put:
masked_image = image.copy() masked_image.paste((0, 0, 0), (96, 32, 288, 192))
condition = Condition("fill", masked_image)
seed_everything() result_img = generate( pipe, prompt="The mona lisa is wearing a A red hat on the head.", conditions=[condition], ).images[0]
concat_image = Image.new("RGB", (1536, 512)) concat_image.paste(image, (0, 0)) concat_image.paste(condition.condition, (512, 0)) concat_image.paste(result_img, (1024, 0)) concat_image
after that cell make anew cell and put:
result_img.save(r"H:\GITHUB\OMINI CONTROL\output\output1.jpg")
thats all