Google's New Whisk AI Use Imgae Prompts Instead of Texts: All Details
Google Whisk AI
By Himani Jha | Updated Dec 18, 2024, 4:49 PM IST
Google Labs has introduced a groundbreaking new AI image generator called "Whisk." Unlike traditional AI image generators that rely solely on text prompts, Whisk allows users to input images directly, opening up new avenues for creative expression and visual exploration. Check out all details about Google's new Whisk AI here.
Google has unveiled a new AI tool named Whisk that allows you to create images by using other images as prompts rather than needing an extensive text prompt. With Whisk, you can provide images to indicate what you desire in terms of the subject, the scene, and the style of your AI-created image, and you can supply Whisk with several images for each of these three categories.
If you lack images, you can select a dice icon to allow Google to provide some photos for the prompts (although those images seem to be AI-generated as well). At the end of the process, you can input text into a box if you'd like to provide additional details regarding the image you seek, though it is not mandatory.
The Whisk will subsequently create images and a text prompt corresponding to each image. If you are pleased with the results, you can favourite or download the image, or you can enhance an image by adding more text in the text box or by clicking the image and modifying the text prompt.
Also Read: Best Tablets in 2024 From Apple, Samsung and More
In a blog entry, Google emphasizes that Whisk is intended for “quick visual exploration, rather than exact pixel adjustments.” The company also states that Whisk might "fall short," which is why it allows users to modify the underlying prompts.
In the brief moments I've utilized the tool while composing this story, it's been enjoyable to experiment with. Generating images takes a few seconds, which can be frustrating, and although the images have been somewhat odd, I've found it enjoyable to iterate on everything I've created.
Also Read: Best Cameras in 2024 for Stunning Photography and Videography
Google states that Whisk employs the “latest” version of its Imagen 3 image generation model, which was revealed today. Google unveiled Veo 2, the latest version of its video generation model, which the company claims comprehends “the distinctive language of cinematography” and imagines elements like additional fingers “more infrequently” compared to other models (one of those models is likely OpenAI’s Sora). Veo 2 will debut with Google's VideoFX, available on the Google Labs waitlist, and will be extended to YouTube Shorts and other products next year.
If you lack images, you can select a dice icon to allow Google to provide some photos for the prompts (although those images seem to be AI-generated as well). At the end of the process, you can input text into a box if you'd like to provide additional details regarding the image you seek, though it is not mandatory.
The Whisk will subsequently create images and a text prompt corresponding to each image. If you are pleased with the results, you can favourite or download the image, or you can enhance an image by adding more text in the text box or by clicking the image and modifying the text prompt.
Also Read: Best Tablets in 2024 From Apple, Samsung and More
In a blog entry, Google emphasizes that Whisk is intended for “quick visual exploration, rather than exact pixel adjustments.” The company also states that Whisk might "fall short," which is why it allows users to modify the underlying prompts.
In the brief moments I've utilized the tool while composing this story, it's been enjoyable to experiment with. Generating images takes a few seconds, which can be frustrating, and although the images have been somewhat odd, I've found it enjoyable to iterate on everything I've created.
Also Read: Best Cameras in 2024 for Stunning Photography and Videography
Google states that Whisk employs the “latest” version of its Imagen 3 image generation model, which was revealed today. Google unveiled Veo 2, the latest version of its video generation model, which the company claims comprehends “the distinctive language of cinematography” and imagines elements like additional fingers “more infrequently” compared to other models (one of those models is likely OpenAI’s Sora). Veo 2 will debut with Google's VideoFX, available on the Google Labs waitlist, and will be extended to YouTube Shorts and other products next year.