Hi all, We are looking to extend IoTDB-AINode's support for large-scale pre-trained time series foundation models. To achieve this, this PR integrates the Moirai2 model as a built-in forecasting option.
Adapted from the original uni2ts repository, we have extracted only the necessary modules required for inference. These have been refactored under the AINode package structure, with training-related and non-essential components removed to streamline the implementation. You can review the changes here: https://github.com/apache/iotdb/pull/17056 Best regards, -------------------- Yuan Tian
