The batch input file to be uploaded should be in .jsonl format, with each line detailing an API inference request.Each request must have a unique custom_id to identify inference results in the output file after batch processing. The parameters within the body field of each line are used as the actual inference request parameters for the endpoint.
All requests within a single batch JSONL file must target the same model. Do not mix requests for different models in one batch.
Here is an example of an input file containing two requests: