Sarvam targets edge AI with compact models

Sarvam, an Indian company specializing in artificial intelligence, is working on extremely small AI models designed to run on resource-constrained devices. The goal is to bring AI to feature phones, cars, and smart glasses, opening up new possibilities for the use of artificial intelligence in different contexts.

Lightweight models for low-power devices

The main feature of these models is their size: they occupy only a few megabytes of space. This makes them suitable for devices with limited memory capacity and reduced computing power. Furthermore, the models are designed to work offline, eliminating the need for a constant internet connection. This is particularly useful in areas with poor connectivity or for applications that require data privacy and security.

For those evaluating on-premise deployments, there are trade-offs to consider. AI-RADAR offers analytical frameworks on /llm-onpremise to evaluate these aspects.