AI can now generate art: A look at popular tools you can use to experience this

Artificial intelligence, or AI, is now actively used to help create images or works of art. And the art created by artificial intelligence has become so good that such works even win prizes at art fairs. For example, in September 2022, the New York Times reported how a 39-year-old man won the top prize in the Colorado State Fair’s fine art competition for an artwork created using artificial intelligence (AI).

Thanks to the development of artificial intelligence, such image generators are now very accurate. You can give them a very specific and weird description like “a man riding a pink horse in the snow with the Taj Mahal in the background” and get new results. This is possible by training the AI ​​on a large set of images with associated meanings pulled from all over the internet. While some of the AI ​​models, like the DALL-E Mini, are open source, meaning their code is available for developers to implement into their own applications, others, like the Midjourney, are proprietary. If you want to learn how to create your own images, here are some of the most popular tools you can use right now.

Five AI image generators you can try

DALL-E 2

DALL·E toyota prius in a cyberpunk themed city DALL-E 2 output from the text description I created “Toyota Prius in a cyberpunk themed city”.

Also Read :  Dimensity 9200 tops current mobile GPUs in early GFXBench test, even the Apple A16

The original DALL-E, created by OpenAI, was released in January 2021 and was replaced this year by DALL-E 2. The latest version generates 4x higher resolution images and maintains significantly higher accuracy. Although the image generator was initially limited to 200 beta testers, it was made available to everyone in September. It uses a dataset of text-image pairs to generate realistic images and art from a natural language description. After signing up, you can try 50 prompts per month.

DALL-E and its sequel DALL-E 2 are built on a customized version of GPT-3 (Generative Pretrained Transformer 3). This is a language-based machine learning model that uses deep learning to generate human-like text. The original GPT-3 was used for text, but the adapted version used by DALL-E is used to generate images using a technique called diffusion modeling.

Halfway

Midjourney is an independent research lab currently working on an AI image generator similar to the ones on this list, with the goal of expanding the imagination of the human species. The AI ​​image generator is currently proprietary. But it can be accessed through the Discord bot. To do this, of course, you need to have a Discord account, and then you’ll need to join Midjourney’s Discord server. The outputs here seem to be mostly artistic, unlike DALL-E 2 and Stable Diffusion, which seem to focus on photorealistic images based on your input. You can create your first 25 images for free, but you can extend the limit to 200 with a basic membership subscription.

Also Read :  Northland’s new mobile healthcare service gets underway in Mid and Far North

StarryAI

StarryAI has become one of the most user-friendly AI image models on this list thanks to its easy-to-use interface. This requires a slightly different approach as it allows you to choose the output type you want based on presets. Altair’s original setting produces dreamlike images closer to abstract art, while Orion’s priority is realism, and Argo’s best for artistic and product rendering.

StarryAI also lets you upload a custom image to base your art on if you have something special. Another unique thing this AI model has to offer is that it has dedicated apps for Android and iOS so you can create and share art on the fly. Note that you can only generate images twice a day and if you want to generate more you will need to purchase additional credits.

Craiyon (formerly DALL-E Mini)

dall and mini craiyon Craiyon is the largest, most powerful open source version of AI DALL-E

Formerly called DALL-E Mini, Craiyon is the largest and most powerful open source version of AI DALL-E. Craiyon was first launched in April 2022. The AI ​​image model is the brainchild of Boris Deim, a machine learning engineer based in Houston.

Also Read :  Commerce Issues Guidance on Recent Advanced Computing and Semiconductor Manufacturing Controls for China | Pillsbury - Global Trade & Sanctions Law

DALL-E Mini is not affiliated with OpenAI and was renamed Craiyon at OpenAI’s insistence. Soon after, it experienced a huge surge in popularity due to its ability to create cartoonish, meme-worthy images. The output is not as realistic as some of the others on the list and has a lot of distortion.

Stable diffusion

stable diffusion dream study Stable diffusion is seen in action here.

Stable Diffusion is another text transformation model released in collaboration between Stability AI, CompVis LMU and Runway. It is supported by EleutherAI and LAION. While you’ll notice some distortion, as with every generator on this list, Stable Diffusion’s outputs are significantly more realistic than some of the others here. The front end of the image generator is DreamStudio, which you can use right now.

Image: Not currently publicly available

Google has been working on its own AI image model called Imagen. Although its output is similar to DALL-E 2 and Stable Diffusion, it has not been made available to the public. This is not likely to change anytime soon. The company has also developed a video version of the AI ​​model based on the photo version, which works by zooming in and “predicting” additional frames.



Source

Leave a Reply

Your email address will not be published.

Related Articles

Back to top button