`` Diffusion Bee '' that can easily execute AI `` Stable Diffusion '' that can generate your favorite images in the macOS environment has appeared, so I tried using it

AI '

Stable Diffusion ', which generates an image according to the input text, is attracting a lot of attention because it can easily output detailed illustrations and photos. Stable Diffusion was developed for PCs equipped with NVIDIA GPUs, but methods of running it on machines with Intel CPUs and Macs with Apple Silicon are being devised one after another. Newly, Mr. Divam Gupta , who is researching AI and VR at Meta, has released a Stable Diffusion execution environment ' Diffusion Bee ' with a UI that runs on Macs with Apple Silicon, so I actually installed it on an iMac with an M1 chip . I tried using it.

GitHub - divamgupta/diffusionbee-stable-diffusion-ui: Diffusion Bee is the easiest way to run Stable Diffusion locally on your M1 Mac. Comes with a one-click installer. No dependencies or technical knowledge needed.

To install Diffusion Bee, go to the distribution page and click 'Download for MacOS' to download the DMG file. The capacity was 561MB.

Then double-click the downloaded DMG file.

When the following screen is displayed, drag and drop the icon on the left onto the icon on the right. The installation is now complete.

After installation is complete, double-click 'DiffusionBee' in the 'Applications' folder to start it. Of course, you can also launch it from Launchpad or from Spotlight search.

Click 'Open' when the following warning is displayed.

The model data will start downloading at the first startup, so wait for a while.

Once the model data has been downloaded, preparations for image generation are complete. After that, enter the description of the image you want to generate in the input field and click 'Generate' to start image generation.

This time,

referring to Stable Diffusion's prompt sharing site ' Lexica ', 'portrait of a teenage girl with bangs, brown hair and bangs, round silver glasses with thin rims, wearing an oversized sweater, dramatic lighting, anime illustration by Greg rutkowski , yoji shinkawa, 4k, digital art, concept art, trending on artstation, anime, featured on pixiv (portrait of teenage girl with bangs, brown hair with bangs, wearing silver rimmed glasses, oversized sweater, dramatic Light source, anime illustration drawn by Greg Rutkowski, Yoji Shinkawa, 4K, digital art, concept art, trending on ArtStation, anime, popular on pixiv)' and clicked 'Generate'.

The generated image is below. The time it took to generate was about a minute and a half. To save the generated image, click 'Save Image' located at the bottom of the image.

Then, the save dialog is displayed, so you can save it in any location.

The actual size (512 × 512 pixels) of the image generated this time is as follows. I was able to generate a fairly detailed illustration in one and a half minutes.

in Review,   Software,   Art, Posted by log1o_hf