🍪 This website uses cookies

    We use cookies to operate our website, analyze traffic, and support marketing activities where permitted by law.
    Learn more in our Cookie Policy.

    GlossaryFine-Tuning
    Glossary · AI

    What is Fine-Tuning?

    Definition

    Fine-Tuning is the process of further training an AI model on specific data to improve performance for a particular task or domain. Unlike general models trained on broad data, Fine-Tuning customizes models for specialized applications. Organizations use Fine-Tuning to adapt AI models to their specific needs.

    model adaptationspecialized trainingdomain expertisemodel customizationtraining optimizationperformance improvementFine-Tuning
    In short

    Fine-Tuning at a glance.

    Trains model on domain-specific data
    Improves performance for specialized tasks
    Reduces need for large training datasets
    Enables cost-effective model customization

    Making AI Yours

    Fine-Tuning allows organizations to adapt powerful AI models to their specific domain or use case. A general language model Fine-Tuned on your company's documents will understand your industry terminology and context better. Fine-Tuning is less expensive than training a model from scratch and much faster than waiting for foundation model improvements.

    Learn more

    AI learning platform

    See how a modern, AI-native platform builds, delivers and tracks training — all in one place.

    Read the guide

    Fine-Tuning — frequently asked

    Fine-Tuning typically works well with hundreds to thousands of examples. More data generally improves results, but too much can reduce efficiency.

    Fine-Tuning can be completed in hours to days, depending on data size and compute resources. Much faster than training from scratch.

    Some Fine-Tuned models can be continuously updated with new data. However, this requires careful management to avoid performance degradation.

    From definition to done.

    See AI learning platform in action — turn your knowledge into training, built and tracked with AI.