fail to train document extractor

Question

I tried to train a custom document extractor, using the minimum set (3 training + 3 test) for template-based training. I've retried for 3 times, all failed with ...

{
  "name": "projects/***/locations/us/operations/***",
  "done": true,
  "result": "error",
  "response": {},
  "metadata": {
    "@type": "type.googleapis.com/google.cloud.documentai.uiv1beta3.TrainProcessorVersionMetadata",
    "commonMetadata": {
      "state": "FAILED",
      "createTime": "2024-02-22T01:11:13.385469Z",
      "updateTime": "2024-02-22T01:15:19.807761Z",
      "resource": "projects/***/locations/us/processors/***/processorVersions/***"
    },
    "trainingDatasetValidation": {},
    "testDatasetValidation": {}
  },
  "error": {
    "code": 13,
    "message": "Internal error encountered.",
    "details": []
  }
}

I've successfully trained other document extractors in the same project. This is the first time I try to read tables (parent/child fields) in Chinese. The labelling should work fine.

labelling