Mistral AI team – August 7, 2024
We’re introducing advancements in building software using language models. These models serve as a bridge between knowledge sources and user interfaces, making it crucial to improve quality, reduce latency, and prototype quickly.
Simpler model customization allows developers to tailor our flagship and specialist models on La Plateforme to their specific applications. This can be done using a base prompt, few-shot prompting, or fine-tuning with your own dataset. Our science team developed techniques for strong reference models, so you can expect similar performance from your fine-tuned models.
Model customization enables developers to integrate generative AI capabilities into their application with domain knowledge, context, or tone. We’re excited to see what groundbreaking applications will be built using this feature. Check out our fine-tuning documentation and try model customization on La Plateforme.
We’re also introducing an early version of Agents, which wraps models with additional context and instruction for exposure on Le Chat or API. Agents help create custom behavior and workflows with simple 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.
We’re working on connecting Agents to tools and data sources and welcome your feedback. Additionally, we’re releasing a stable version of our client SDK, mistralai 1.0, available for Python and Typescript. This update improves usability and consistency. Learn more about the new SDK and check out the migration guide.
Source: https://mistral.ai/news/build-tweak-repeat/