Skip to main content

​
About the Provider

Pruna AI is an AI optimization company focused on making state-of-the-art image generation fast, affordable, and accessible. Their P-Image model family is designed to deliver professional-quality image generation with sub-second inference speeds, strong prompt adherence, and support for fine-tuning and LoRA-based customization.

​
Model Quickstart

This section helps you quickly get started with the p-image-lora model on the Qubrid AI inferencing platform. To use this model, you need:
  • A valid Qubrid API key
  • Access to the Qubrid inference API
  • Basic knowledge of making API requests in your preferred language
Once authenticated with your API key, you can send inference requests to the p-image-lora model and receive responses based on your input prompts. Below are example placeholders showing how the model can be accessed using different programming environments.
You can choose the one that best fits your workflow. 
import requests
import json

url = "https://platform.qubrid.com/v1/images/generations"
headers = {
    "Authorization": "Bearer QUBRID_API_KEY",
    "Content-Type": "application/json"
}

data = {
  "model": "p-image-lora",
  "prompt": "cinematic shot of a lone astronaut standing on a desolate alien planet, glowing orange sunset sky, dust storms swirling, dramatic lighting, ultra-wide lens composition, movie still aesthetic, realistic space suit details, volumetric atmosphere, 8k sci-fi film scene",
  "lora_weights": None,
  "lora_scale": 1,
  "hf_api_token": None,
  "aspect_ratio": "16:9",
  "width": 1024,
  "height": 1024,
  "seed": 0,
  "prompt_upsampling": False,
  "disable_safety_checker": False,
  "response_format": "url"
}

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

if response.status_code == 200:
    with open("generated_image.png", "wb") as f:
        f.write(response.content)
        print("Image saved to generated_image.png")
else:
    print(f"Error: {response.status_code}")
    print(response.text)

​
API Generated Response

  • Input Image
Edit Prompt: Add a rainbow over the mountains
  • Output Image

​
Model Overview

P-Image LoRA is a variant of the P-Image model developed by Pruna AI, extending the base model with LoRA (Low-Rank Adaptation) support for fine-tuned style control.
  • It enables custom LoRA weights to be applied at inference time, allowing domain-specific visual style adaptation without full model retraining.
  • The model supports optional prompt upsampling for enhanced prompt interpretation and more detailed image outputs.
  • P-Image LoRA is suitable for production workflows requiring both speed and style customization through lightweight fine-tuning.

​
Model at a Glance

FeatureDetails
Model IDp-image-lora
ProviderPruna AI
Model TypeText-to-Image Generation with LoRA Support
ArchitectureOptimized latent diffusion model with LoRA adapter support
Default Resolution1024 × 1024
Output TypeImages generated from text prompts with optional LoRA style control

​
When to use?

You should consider using P-Image LoRA if:
  • You need high-quality text-to-image generation with custom style control via LoRA weights
  • Your application requires domain-specific visual adaptation without full model retraining
  • You want optional prompt upsampling for more detailed and accurate image outputs
  • You need fast, scalable image generation with lightweight fine-tuning support

​
Inference Parameters

Parameter NameTypeDefaultDescription
LoRA WeightsstringnullPath or identifier of the LoRA weights to apply at inference.
LoRA Scalenumber1Strength of the LoRA adapter effect on the output image.
HF API TokenstringnullHugging Face API token required for accessing private LoRA weights.
Aspect Ratiostring16:9Output image aspect ratio.
Widthnumber1024Image width in pixels.
Heightnumber1024Image height in pixels.
Seednumber0Random seed for reproducibility.
Prompt UpsamplingbooleanfalseEnables prompt enhancement for more detailed image outputs.
Disable Safety CheckerbooleanfalseDisables the built-in safety filter if enabled.
Response FormatstringurlFormat of the response — url or b64_json.

​
Key Features

  • LoRA Style Control: Supports custom LoRA weights at inference time for domain-specific visual style adaptation without full model retraining.
  • Prompt Upsampling: Optional prompt enhancement for more detailed and accurate interpretation of complex text prompts.
  • Hugging Face Integration: Accepts Hugging Face API tokens for seamless access to private or community LoRA weights.
  • Sub-Second Inference: Maintains the fast inference speeds of the P-Image base model with added LoRA customization.
  • Flexible Resolution Control: Configurable width, height, and aspect ratio for diverse output requirements.

​
Summary

P-Image LoRA is a fast, customizable text-to-image model built by Pruna AI for production-scale image generation with style control.
  • It extends the P-Image base model with LoRA adapter support for lightweight fine-tuning at inference time.
  • The model supports Hugging Face integration for accessing private or community LoRA weights.
  • It is suitable for developers and enterprises requiring scalable image generation with domain-specific visual customization.