custom_id
(unique id) and body
fieldsbatch_input_data.jsonl
, making sure custom_id
is unique across rows.
UI
, firectl
or HTTP API
.
Create Dataset
and follow the wizard.State | Description |
---|---|
VALIDATING | The input dataset is being validated to ensure it meets format requirements and constraints |
PENDING | The job is queued and waiting for available resources to begin processing |
RUNNING | The batch job is actively processing requests from the input dataset |
COMPLETED | All requests have been successfully processed and results are available in the output dataset |
FAILED | The job encountered an unrecoverable error. Check the job status message for details |
EXPIRED | The job exceeded the 24-hour time limit. Any completed requests up to that point are saved to the output dataset |
File Type | Description |
---|---|
Results File | Contains successful inference responses in JSONL format, with each line matching the custom_id from your input |
Error File | Contains any error details for requests that failed processing, and the original custom_id for debugging |
max_tokens
limits to optimize processing time.custom_id
values for easier result tracking.Python builder SDK
references
HTTP API
references
firectl
references