​
Creating a Synthetic Dataset

Follow these simple steps to generate your synthetic dataset using Sinkove’s platform:

​
Step 1: Select Model

Navigate to the Generator tab and choose a suitable generative model from the dropdown menu. Our models are optimized for different types of medical imaging scenarios.

Different models may specialize in specific imaging modalities or medical conditions. Choose the one that best fits your research requirements.

​
Step 2: Configure Dataset

Fill in the provided fields with the details of the medical images you wish to generate:

  • Prompts/Descriptions: Enter relevant medical descriptions (e.g., “Severe cardiomegaly”, “Normal chest X-ray”)
  • Dataset Size: Currently limited to approximately 1000 images
  • Additional Parameters: Configure any model-specific settings as needed

Use specific, clinical terminology in your prompts for the most accurate synthetic image generation.

​
Step 3: Generate Data

You have two options for data generation:

​
Preview Generation

  • Click Generate sample to preview a single synthetic image
  • Review the quality and adjust parameters if needed
  • This helps ensure the output matches your expectations

​
Full Dataset Generation

  • Set your desired dataset size (up to ~1000 images)
  • Click Generate dataset to initiate full dataset generation
  • Monitor the progress in your dashboard

Dataset generation may take several minutes to hours depending on the size and complexity of your request.

​
Downloading Your Dataset

Once your dataset is ready, use our Python SDK for quick and efficient downloading:

import sinkove as sk
import time

# Initialize the client with your API key
my_sk = sk.client("your_API_KEY")

# Replace with your actual dataset ID
dataset = my_sk.dataset("dataset_id_here")

# Wait for dataset to be ready
while not dataset.ready():
    print("Dataset not ready, waiting...")
    time.sleep(10)

# Download the dataset
dataset.save("./dataset.zip")
print("Dataset downloaded successfully!")

Next Steps

​
Need Help?

If you encounter any issues or have questions: