Skip to main content

Model Training

Fine-tune and pretrain your own LLMs and other generative AI models

Fine-tune an open source LLM or build custom LLMs trained on your enterprise data with Mosaic AI Model Training. Custom models built with Model Training are faster, produce higher-quality results that are more domain-specific, and have up to 10x lower costs than proprietary LLMs.

A complex image with various elements, including text, diagrams, and charts.

Highly accurate

Fine-tuning an open source LLM or building a new LLM with enterprise data leads to a greater semantic understanding of the business and delivers highly accurate responses. Because Mosaic AI Model Training is natively available in Databricks, organizations can easily and securely fine-tune or build models without moving their data. This also ensures governance, auditability, traceability and monitoring to ensure models are used in the right way and are providing the right responses. The result is higher-quality and accurate results that are specific to the business context.

Pretraining compute plane graphic image

Effortless scale

A key element to high-performance LLM training is scalability, which requires fast, low-latency networking and access to the highest-performing GPUs. Using Mosaic AI Model Training automatically gives you access to both NVIDIA InfiniBand networking and NVIDIA H100 Tensor Core GPUs, the highest-performing NVIDIA GPUs, which give unprecedented performance and scalability compared to previous hardware generations. This lets you scale to train large (>70 billion-parameter) models easily and complete training runs in hours and days.

Pretraining shows it can train a Stable Diffusion model for 10x less cost

Cost-effective

Mosaic AI Model Training can fine-tune smaller open source GenAI LLMs to produce highly efficient models that can be served up to 5x more cost-effectively than larger proprietary LLMs. Additionally, you can build new LLMs from scratch using an optimized software stack that makes training LLMs cost-effective. A combination of system-level optimizations, tuned parallelism strategies and model training science results in a 10x lower cost of training.

model training architecture

Secure and compliant

For most organizations, security is paramount, and they can’t afford to have their employees send their data to a third-party API and risk having the data leaked or used to train a public model. Mosaic AI Model Training ensures that this can never happen, because organizations will build their own LLM where they maintain complete control and ownership over both the data and the model. Everything remains encrypted by default, including traffic and all training data. This ensures you have complete data privacy and full model ownership, meeting any regulatory compliance.

Ready to get started?