sparklearningstudiollc/nikujkakdiya-new-model - AI Image Tools Tool

Overview

sparklearningstudiollc/nikujkakdiya-new-model is an image-generation model hosted on Replicate that exposes text-to-image, image-to-image, and inpainting modes through the platform's API. The model emphasizes configurability: users can set prompt strength for guidance blending, supply custom canvas dimensions or aspect ratios, apply LoRA (Low-Rank Adaptation) weights, and tune output settings such as sampling steps, guidance scale, seed, and number of images. These controls make it suitable for both creative generation and iterative image editing workflows. Because the model is delivered via Replicate, it is accessible through the standard Replicate prediction API and integrates with common tooling and deployment patterns on that platform. According to the model page on Replicate (https://replicate.com/sparklearningstudiollc/nikujkakdiya-new-model/versions/74805ca9e286564b99e477d649945564ed7ad82338ac98f7d44af2e3bd7c1746/api), the API supports supplying initial images and masks for inpainting, specifying LoRA identifiers and weights to influence style, and output controls for batch size and image dimensions. Pricing and rate limits are not published on the model page; those details depend on Replicate account settings or the model author's configuration.

Key Features

• Text-to-image generation with tunable guidance and sampling step controls
• Image-to-image mode accepting an initial image for restyling or refinement
• Inpainting via mask input to selectively edit regions of an image
• Custom dimensions and aspect ratio settings for exact output sizes
• LoRA integration with configurable weight to apply external style models

Code Examples

Python

import os
import requests

# Uses Replicate's prediction API. Set REPLICATE_API_TOKEN in environment.
API_TOKEN = os.environ.get('REPLICATE_API_TOKEN')
URL = 'https://api.replicate.com/v1/predictions'

headers = {
    'Authorization': f'Token {API_TOKEN}',
    'Content-Type': 'application/json'
}

# Example: text-to-image generation with custom dimensions and LoRA weight
payload = {
    'version': '74805ca9e286564b99e477d649945564ed7ad82338ac98f7d44af2e3bd7c1746',
    'input': {
        'prompt': 'A serene mountain lake at sunrise, ultra-detailed, cinematic lighting',
        'negative_prompt': 'lowres, deformed, watermark',
        'width': 1024,
        'height': 768,
        'prompt_strength': 0.9,
        'num_outputs': 2,
        'seed': 42,
        'steps': 28,
        'guidance_scale': 7.5,
        'lora': 'username/lora-model-name',
        'lora_weight': 0.6
    }
}

resp = requests.post(URL, headers=headers, json=payload)
resp.raise_for_status()
print(resp.json())

Curl

curl -s -X POST https://api.replicate.com/v1/predictions \
  -H "Authorization: Token $REPLICATE_API_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "version": "74805ca9e286564b99e477d649945564ed7ad82338ac98f7d44af2e3bd7c1746",
    "input": {
      "prompt": "Portrait of a young woman in renaissance style, warm color palette",
      "width": 768,
      "height": 1024,
      "steps": 30,
      "guidance_scale": 8,
      "num_outputs": 1
    }
  }'

# For image-to-image or inpainting, include "init_image" (URL/base64) and optional "mask" keys in input.

Javascript

const fetch = require('node-fetch');

// Example using Node.js fetch. Set process.env.REPLICATE_API_TOKEN.
const API_TOKEN = process.env.REPLICATE_API_TOKEN;
const url = 'https://api.replicate.com/v1/predictions';

const body = {
  version: '74805ca9e286564b99e477d649945564ed7ad82338ac98f7d44af2e3bd7c1746',
  input: {
    prompt: 'A futuristic cityscape at night, neon reflections, wide-angle',
    negative_prompt: 'blurry, text, watermark',
    width: 1280,
    height: 720,
    steps: 25,
    guidance_scale: 6.5,
    num_outputs: 1
  }
};

fetch(url, {
  method: 'POST',
  headers: {
    'Authorization': `Token ${API_TOKEN}`,
    'Content-Type': 'application/json'
  },
  body: JSON.stringify(body)
})
  .then(res => res.json())
  .then(data => console.log(data))
  .catch(err => console.error(err));

API Overview

• Authentication: API key
• Base URL: https://replicate.com/sparklearningstudiollc/nikujkakdiya-new-model/versions/74805ca9e286564b99e477d649945564ed7ad82338ac98f7d44af2e3bd7c1746/api

Key Information

• Category: Image Tools
• Type: AI Image Tools Tool