LCM-LoRAs for blazing fast generations

You can use LCM-LoRAs via our Image Gen APIs, which are documented in detail at Image Gen API Docs

The following guidelines must be adhered to ensure high quality of output:

1. cfg_scale must be set to a small number between 1-2 inclusive.
2. You must use the octoai:lcm_sdxl checkpoint, which is available in the OctoAI Asset Library
3. You must use the LCM sampler

Note: if your API request does not follow the above guidelines, we will automatically force your API parameters to follow the acceptable values.

Typically, 4 steps should be sufficient for high-quality output.

Here is a curl example for text2img with LCM-LoRA and SDXL:

curl -H 'Content-Type: application/json' -H "Authorization: Bearer $OCTOAI_TOKEN" -X POST "https://image.octoai.run/generate/sdxl" \
    -d '{
        "prompt": "A puppy playing in a pond",
        "negative_prompt": "ugly, tiling, poorly drawn hands, poorly drawn feet, poorly drawn face, out of frame, extra limbs, disfigured, deformed, body out of frame, blurry, bad anatomy, blurred, watermark, grainy, signature, cut off, draft",
        "steps": 4,
        "seed": 2748252853,
        "sampler": "LCM",
        "checkpoint": "octoai:lcm_sdxl",
        "cfg_scale": 2
    }'

And here's a Python example:

import requests
import os
import base64
import io
import PIL.Image


def _process_test(url):
    
    OCTOAI_TOKEN = os.environ.get("OCTOAI_TOKEN")

    payload = {
        "prompt": "A puppy playing in a pond",
        "negative_prompt": "ugly, tiling, poorly drawn hands, poorly drawn feet, poorly drawn face, out of frame, extra limbs, disfigured, deformed, body out of frame, blurry, bad anatomy, blurred, watermark, grainy, signature, cut off, draft",
        "steps": 4,
        "seed": 2748252853,
        "sampler": "LCM",
        "checkpoint": "octoai:lcm_sdxl",
        "cfg_scale": 2
    }

    headers = {
        "Authorization": f"Bearer {OCTOAI_TOKEN}",
        "Content-Type": "application/json",
    }

    response = requests.post(url, headers=headers, json=payload)

    if response.status_code != 200:
        print(response.text)

    img_list = response.json()["images"]

    for i, img_info in enumerate(img_list):
        img_bytes = base64.b64decode(img_info["image_b64"])
        img = PIL.Image.open(io.BytesIO(img_bytes))
        img.load()
        img.save(f"result_image{i}.jpg")


if __name__ == "__main__":
    _process_test("https://image.octoai.run/generate/sdxl")