1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
import requests
import base64
# Use this function to convert an image file from the filesystem to base64
def image_file_to_base64(image_path):
with open(image_path, 'rb') as f:
image_data = f.read()
return base64.b64encode(image_data).decode('utf-8')
# Use this function to fetch an image from a URL and convert it to base64
def image_url_to_base64(image_url):
response = requests.get(image_url)
image_data = response.content
return base64.b64encode(image_data).decode('utf-8')
api_key = "YOUR_API_KEY"
url = "https://api.segmind.com/v1/sdxl1.0-realdream-lightning"
# Request payload
data = {
"prompt": "girl in a hat, in a dress Russian patterns, freckles, cornfield, nature background, beautiful landscape, retro style, fashion, magazine cover, professional quality, aesthetics, gentle, beautiful, realism,high detail,high quality, 64k, high resolution, professional photo",
"negative_prompt": "blemishes,(worst quality, low quality, illustration, 3d, 2d, painting, cartoons, sketch), open mouth",
"samples": 1,
"scheduler": "DPM++ SDE",
"num_inference_steps": 8,
"guidance_scale": 1.2,
"seed": 968875,
"img_width": 768,
"img_height": 1152,
"base64": False
}
headers = {'x-api-key': api_key}
response = requests.post(url, json=data, headers=headers)
print(response.content) # The response is the generated imageAttributes
Prompt to render
blemishes,(worst quality, low quality, illustration, 3d, 2d, painting, cartoons, sketch), open mouth
Number of samples to generate.
min : 1,
max : 4
Type of scheduler.
Allowed values:
Number of denoising steps.
min : 1,
max : 100
Scale for classifier-free guidance
min : 1,
max : 25
Seed for image generation.
min : -1,
max : 999999999999999
Can only be 1024 for SDXL
Allowed values:
Can only be 1024 for SDXL
Allowed values:
Base64 encoding of the output image.
To keep track of your credit usage, you can inspect the response headers of each API call. The x-remaining-credits property will indicate the number of remaining credits in your account. Ensure you monitor this value to avoid any disruptions in your API usage.
RealDream Lightning
Real Dream is a sophisticated image generation model utilizing SDXL Lightning architecture. It creates incredibly realistic images from textual prompts. With the ability to excellently generate human portraits from the user's descriptive text, the model stands out in the level of detail and realism it outputs.
-
Capabilities: Real Dream is skilled at producing high-quality, realistic images. It is particularly proficient in creating detailed and lifelike portraits.
-
Creators: The model was developed by a user by the pseudonym "sinatra" and is hosted on Civitai.
-
Training Data: The training data consists of a diverse assortment of high-resolution images with a focus on realism, especially various human features.
-
Technical Architecture: Real Dream uses the SDXL Lightning architecture, known for its high-quality image generating capabilities.
-
Strengths: The model generates highly realistic images, capturing intricate textures and features. It is also versatile in producing diverse human images.
How to use the model?
Enter a detailed description of the image in the "Prompt" text box. Put a seed value in the "seed" box for reproducibility or enable the "Randomize Seed" toggle for randomness. Click on "Advanced Parameters" for more customization options.In the "Negative Prompt" field, indicate any unwanted elements in the image. Select a scheduler option from the dropdown, e.g., "DPM++ SDE." Indicate the number of steps for the generation process. Fine-tune image adherence to your text prompt with the guidance scale. et the output image dimensions using "img_width" and "img_height."Click "Generate" to create your image.
