'style_preset' doesn't work

[geosem42]

I'm trying now the new 'style_preset' but it doesn't work, passing a value doesn't do anything.

However, specifying the style manually in the prompt works.

Even passing the value manually doesn't do anything style_preset: 'anime'

const { res: result, images } = await generateAsync({
        prompt,
        apiKey,
        height: height,
        width: width,
        cfg_scale: 7,
        steps: 50,
        num_samples: 1,
        engine: 'stable-diffusion-xl-beta-v2-2-2',
        style_preset // or style_preset: style_preset
})

[brianfitzgerald]

Hi Geosemaan, do you mind sharing the details of your generateAsync function, or what the body and URL of your request look like? Just tried to reproduce your request with the following cURL:

curl --request POST \
  --url https://api.stability.ai/v1/generation/stable-diffusion-xl-beta-v2-2-2/text-to-image \
  --header 'Accept: image/png' \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '{
    "text_prompts": [
        {
            "text": "dog on a swing"
        }
    ],
    "cfg_scale": 7,
    "steps": 50,
    "sampler": "DDIM",
    "style_preset": "anime"
}'

and got the expected response.

[geosem42]

Hello Brian, I'm using stability-client package: https://www.npmjs.com/package/stability-client

It looks like it hasn't been updated yet. I tried patching it locally for testing, it didn't work.

[brianfitzgerald]

Ah! That makes sense. That is a third party client that is not supported by Stability. I would open an issue / PR on their repo, or in the meantime, you can reference our TypeScript code examples here: https://api.stability.ai/docs#tag/v1generation/operation/textToImage

[brianfitzgerald]

Going to close this issue since we've determined a root cause, but feel free to ping me or raise again if you run into other issues.