MaxCompute - MaxFrame AI Function released
Apr 09 2025
MaxComputeMaxCompute introduces the MaxFrame AI Function feature, which can be used to call large language models (LLMs) for processing large amounts of data offline.
Content
Target customers: data development engineers and data scientists. New features and specifications: MaxFrame AI Function integrates LLMs such as Qwen2.5 and Deepseek-R1-Distill-Qwen. You can use simple programming interfaces to directly call the LLMs to process large amounts of data in MaxCompute tables offline. The LLMs are hosted offline in MaxCompute. You do not need to consider the download and distribution of models, and the maximum concurrency of API calls. Therefore, MaxFrame jobs that call MaxFrame AI Function can make full use of the massive computing resources of MaxCompute and complete text processing tasks based on LLM inference with high token throughput and concurrency.
