test-yash-model-4-new-2 - AI Image Tools Tool

Overview

test-yash-model-4-new-2 is a custom diffusion-based text-to-image model focused on generating fashion design concepts from natural-language prompts. Hosted on Replicate, the model exposes API parameters that let developers control generation factors such as prompt strength, aspect ratio, model selection, and output format to produce tailored garment visuals and concept art. According to the model's API reference on Replicate (https://replicate.com/stoodioai/test-yash-model-4-new-2/api/schema), the endpoint schema is documented so integrators can programmatically adjust synthesis parameters. The model is positioned for designers, fashion tech teams, and prototyping workflows that need quick visual variations rather than production-ready photoreal renders. By adjusting prompt strength and aspect ratio, users can trade off between strict adherence to a prompt and stylistic variance, or generate images optimized for specific presentation layouts (e.g., lookbooks or social posts). The API-centric design enables batch generation, different output formats, and model selection where multiple model variants may be available via the same Replicate project page.

Key Features

• Diffusion-based text-to-image generation tailored to fashion and apparel concepts
• Prompt strength control to adjust adherence versus creative variation
• Aspect ratio parameter for outputs formatted to lookbooks, CGD, or social layouts
• Multiple model selection options exposed via the Replicate API schema
• Configurable output formats (e.g., PNG) for direct use in design workflows

Example Usage

Example (python):

import os
import replicate

# Ensure REPLICATE_API_TOKEN is set in your environment
# pip install replicate

client = replicate.Client(api_token=os.environ.get("REPLICATE_API_TOKEN"))

# Get the model by name on Replicate
model = client.models.get("stoodioai/test-yash-model-4-new-2")
# Select a version (the API page lists available versions)
version = model.versions[0]

# Example input payload - parameter names (prompt, prompt_strength, aspect_ratio, output_format, model_choice)
# are based on the model's Replicate schema; consult the API reference for exact parameter options.
input_payload = {
    "prompt": "Futuristic evening gown with asymmetrical sleeves, metallic accents, and flowing chiffon",
    "prompt_strength": 0.8,
    "aspect_ratio": "3:4",
    "output_format": "png",
    "model_choice": "default"
}

# Create a prediction (synchronous). This will return a prediction object with output URLs once complete.
prediction = client.predictions.create(version=version.id, input=input_payload)

# Poll or wait for the prediction to complete (the client may block until done)
# Print resulting image URLs or metadata
print("Prediction status:", prediction.status)
print("Output:", prediction.output)

Key Information

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