Content

Target customers: Customers who use EAS to build LLM-driven applications and services, such as intelligent customer service, content generation, and translation. New features /specifications: BladeLLM is an inference engine developed by PAI. It provides efficient runtime, high-performance operator implementation, and extreme hybrid quantization. PAI-EAS fully integrates BladeLLM to launch the LLM high-performance inference service. It supports the deployment of preset models and custom models as well as advanced options such as model parallelism and speculative sampling. This provides customers with efficient LLM deployment solutions.