image generation
Krea 2 Turbo
Krea 2 TurboKrea 2 is a text-to-image diffusion model that generates images from natural-language text descriptions.
Model details
View repositoryExample usage
Krea 2 Turbo is the fast, post-trained release of Krea’s 12-billion-parameter text-to-image diffusion model. It generates high-quality images from natural-language prompts in as few as eight inference steps, making it well suited for rapid concepting, design exploration, visual production, and creative applications.
Released with open weights, Krea 2 Turbo gives developers and researchers the flexibility to run, customize, and integrate the model into their own workflows. It is designed around strong prompt following, aesthetic diversity, and expressive style control, helping creators move quickly from an idea to a polished visual.
Input
1import os
2import base64
3from openai import OpenAI
4
5MODEL_ID = "" # Replace with the Model ID in Baseten
6ENDPOINT = f"https://model-{MODEL_ID}.api.baseten.co/environments/production/sync/v1"
7
8client = OpenAI(
9 # Add your Baseten API Key as an environment variable
10 api_key = os.getenv("BASETEN_API_KEY"),
11 base_url = ENDPOINT
12)
13
14response = client.images.generate(
15 prompt=(
16 "A red fox sitting in fresh snow at golden hour, "
17 "photorealistic, cinematic lighting"
18 ),
19 size = "1024x1024",
20 n=1,
21 response_format="b64_json",
22)
23
24image_bytes = base64.b64decode(response.data[0].b64_json)
25
26with open("krea-2-turbo-output.png", "wb") as file:
27 file.write(image_bytes)
28
29print("Saved image to krea-2-turbo-output.png")JSON output
1{
2 "created": 1784218289,
3 "data": [
4 {
5 "b64_json": "<base64-encoded image data>",
6 "revised_prompt": "A red fox sitting in fresh snow at golden hour, photorealistic, cinematic lighting",
7 "url": null,
8 "file_path": "/workspace/outputs/id.jpg"
9 }
10 ],
11 "id": "id",
12 "peak_memory_mb": 53126,
13 "inference_time_s": 1.5457863844931126
14}
