August 7, 2024
Mistral AI team
Language models have changed how software is built, acting as a flexible bridge between knowledge sources and user interfaces. However, this also presents challenges to improve quality, reduce latency, and prototype quickly. To address these issues, we are announcing several advancements.
Simpler model customization: Our flagship and specialist models can now be customized using a base prompt, few-shot prompting, or fine-tuning, along with the option to bring your own dataset. This customization process is based on techniques developed by our science team for creating strong reference models, ensuring similar performance from fine-tuned models.
Developers can use model customization to integrate generative AI capabilities into their application, incorporating specific domain knowledge, context, or tone. We expect this feature to unlock groundbreaking applications and are eager to see what developers will build with it.
Alpha release of Agents: We are introducing an early version of Agents that wraps models with additional context and instruction for use on Le Chat or API. Agents enable custom behavior and workflows by following a simple set of instructions and examples. With Mistral Large 2’s advanced reasoning capabilities, you can layer complex workflows with multiple agents that are easy to share within your organization.
Stable version of our client SDK: We have updated the mistralai library to improve usability and consistency, and today we are releasing mistralai 1.0, available for both Python and Typescript. Learn more about our new SDK and check out the migration guide.
Source: https://mistral.ai/news/build-tweak-repeat/?utm_source=perplexity