Skip to content

environment_variable_instructions.md

Latest commit

 

History

History
22 lines (18 loc) · 1.13 KB

environment_variable_instructions.md

File metadata and controls

22 lines (18 loc) · 1.13 KB

Environment Variable Configuration Instructions

This document provides guidance on configuring environment variables for the Gemini Batch API workflow.

Initial Setup

During the initial setup, you do not need to specify a value for the OUTPUT_JSONL_PATH environment variable. Leave it blank or as it is.

Important Notes

  • Location: Do not change the location setting; leave it set to "global".
  • Ensure all other required environment variables are correctly set.

Post-Batch Processing

After the batch job has finished, update the OUTPUT_JSONL_PATH before running the post-processing pipeline.

  1. Locate the output folder in your Google Cloud Storage (GCS) bucket.
  2. The folder will have a name resembling prediction-model-2026-03-19... followed by a timestamp.
  3. Copy the full folder name.
  4. Set the OUTPUT_JSONL_PATH variable to the following format: 
    OUTPUT_JSONL_PATH="jsonl/prediction-model-2026-03-19.../"
    Replace prediction-model-2026-03-19... with the actual folder name from your bucket.

This ensures the post-processing script can access the correct output JSONL file generated by the batch job.