Mistral, a French AI startup, has launched new services and an SDK to allow customers to fine-tune its models. The SDK, called Mistral-Finetune, is optimized for multi-GPU setups but can also scale down to a single Nvidia A100 or H100 GPU for fine-tuning smaller models.
Mistral also offers fine-tuning services through its API for developers and companies who prefer a managed solution. Currently, the services are compatible with Mistral Small and Mistral 7B models.
A custom training service, currently only available to select customers, is also offered, allowing users to fine-tune any Mistral model for an organization’s apps using their data.
By using this site, you agree to allow SPEEDA Edge and our partners to use cookies for analytics and personalization. Visit our privacy policy for more information about our data collection practices.