This post covers a free image to image AI tool to repurpose photos online. The name of the tool is PhotoMaker by TencentART hosted on HuggingFace. The tool allows users to add one or more photos along with a prompt and generates customized photos. It can generate photos in 10 different styles covering Photographic, Fantasy Art, Cinematic, Digital Art, Disney Character, and more.
In terms of technicalities, the model is based on the “PhotoMaker: Customizing Realistic Human Photos via Stacked ID Embedding” paper and contains two main parts: an ID encoder and LoRA weights. It can be adapted to any base model based on SDXL or used in conjunction with other LoRA models. It is hosted on HuggingFace where anyone can visit and use it for free. It takes very little time to generate photos with a resolution of 1024×1024 pixels.
Check out our other posts to learn how to remove watermarks from photos with AI and swap faces in photos and videos using AI.
Repurpose Photos Online using AI
Head over to PhotoMaker on HuggingFace using this direct link. The website has input parameters on the left and an output section on the right. The first input parameter is the image upload. Here you have to upload the photo that you want to repurpose. You can upload multiple photos to get better results in the output.
Then comes the Prompt where you have to define what you want in the output. This tool uses “img” as a trigger. You have to use this word right after the class word that you want to customize. For example, if your photo contains a man and a woman and you want to use the woman in the output then you have to add the trigger word after woman (e.g., woman img).
The next input is photo styles. The tool covers the following photo styles:
- (No style)
- Cinematic
- Disney Character
- Digital Art
- Photographic (Default)
- Fantasy art
- Neonpunk
- Enhance
- Comic book
- Lowpoly
- Line art
By adding the input image(s), prompt with trigger word and the photo style, you can run the tool to get the output. This tool generates two images per run by default that you can change. The images have a download and a share button that you can use to save and share the generated images. The resolution of the images is 1024×1024 px.
The tool comes with Advanced options that let you tweak the image generation model. These options let you configure negative prompts, number of sample steps, style strength, number of output images, guidance scale, and seed. You can configure these options as per your liking. Changing the option may impact the generation time.
You can try it here.
Closing Words
PhotoMaker is a handy tool that lets you easily repurpose images. You can upload a model or person and easily generate variations of that with different clothing, locations, scenarios, etc. This model struggles with Asian male faces and accurately rendering human hands. These particular scenarios might include degraded customization performance.
