Sign in

v1

latestOpenAPI 3.1.02026-07-26106201349.6 KB
Gateway

Create Reinforcement Fine-tuning Job

post/v1/accounts/{account_id}/reinforcementFineTuningJobs

Path parameters

account_idstring required

The Account Id

Query parameters

reinforcementFineTuningJobIdstring

ID of the reinforcement fine-tuning job, a random UUID will be generated if not specified.

Request body

namestring
displayNamestring
createTimestring date-time
completedTimestring date-time

The completed time for the reinforcement fine-tuning job.

datasetstring required

The name of the dataset used for training.

evaluationDatasetstring

The name of a separate dataset to use for evaluation.

evalAutoCarveoutboolean

Whether to auto-carve the dataset for eval.

state'JOB_STATE_UNSPECIFIED' | 'JOB_STATE_CREATING' | 'JOB_STATE_RUNNING' | 'JOB_STATE_COMPLETED' | 'JOB_STATE_FAILED' | 'JOB_STATE_CANCELLED' | 'JOB_STATE_DELETING' | 'JOB_STATE_WRITING_RESULTS' | 'JOB_STATE_VALIDATING' | 'JOB_STATE_DELETING_CLEANING_UP' | 'JOB_STATE_PENDING' | 'JOB_STATE_EXPIRED' | 'JOB_STATE_RE_QUEUEING' | 'JOB_STATE_CREATING_INPUT_DATASET' | 'JOB_STATE_IDLE' | 'JOB_STATE_CANCELLING' | 'JOB_STATE_EARLY_STOPPED' | 'JOB_STATE_PAUSED' | 'JOB_STATE_DELETED' | 'JOB_STATE_ARCHIVED'

JobState represents the state an asynchronous job can be in.

  • JOB_STATE_PAUSED: Job is paused, typically due to account suspension or manual intervention.
  • JOB_STATE_DELETED: Job has been deleted.
  • JOB_STATE_ARCHIVED: User-facing state for jobs whose row is retained post-delete (e.g. RLOR trainers within the checkpoint retention window). The internal row is still in JOB_STATE_DELETED; the gateway translates it to ARCHIVED on public responses.
createdBystring

The email address of the user who initiated this fine-tuning job.

evaluatorstring required

The evaluator resource name to use for RLOR fine-tuning job.

outputStatsstring

The output dataset's aggregated stats for the evaluation job.

chunkSizeinteger

Data chunking for rollout, default size 200, enabled when dataset > 300. Valid range is 1-10,000.

outputMetricsstring
maxInferenceReplicaCountinteger
nodeCountinteger

The number of nodes to use for the fine-tuning job. If not specified, the default is 1.

trainerLogsSignedUrlstring

The signed URL for the trainer logs file (stdout/stderr). Only populated if the account has trainer log reading enabled.

acceleratorSecondsobject

Accelerator seconds used by the job, keyed by accelerator type (e.g., "NVIDIA_H100_80GB"). Updated when job completes or is cancelled.

maxConcurrentRolloutsinteger

Maximum number of concurrent rollouts during the RFT job.

maxConcurrentEvaluationsinteger

Maximum number of concurrent evaluations during the RFT job.

purpose'PURPOSE_UNSPECIFIED' | 'PURPOSE_PILOT'

Scheduling purpose for training jobs and deployments.

encryptionState'ENCRYPTION_STATE_UNSPECIFIED' | 'ENCRYPTION_STATE_PLAINTEXT' | 'ENCRYPTION_STATE_CMEK'

EncryptionState is the authoritative, per-resource CMEK marker: the single source of truth for whether a resource's customer-data artifacts are encrypted. It is stamped at resource creation from the owning account's CMEK config and is immutable for the resource's life, so a resource always reads back with the state it was created under regardless of the account's current flag. The encrypt-on-write / decrypt-on-read decision for the resource's data keys off this field.

  • ENCRYPTION_STATE_UNSPECIFIED: Unstamped/legacy rows that predate CMEK. Treated as PLAINTEXT.
  • ENCRYPTION_STATE_PLAINTEXT: Customer-data artifacts are stored as plaintext (Fireworks-owned, no CMEK).
  • ENCRYPTION_STATE_CMEK: Customer-data artifacts are CMEK-encrypted; reads MUST decrypt (fail-closed).

Response

A successful response.

namestring
displayNamestring
createTimestring date-time
completedTimestring date-time

The completed time for the reinforcement fine-tuning job.

datasetstring required

The name of the dataset used for training.

evaluationDatasetstring

The name of a separate dataset to use for evaluation.

evalAutoCarveoutboolean

Whether to auto-carve the dataset for eval.

state'JOB_STATE_UNSPECIFIED' | 'JOB_STATE_CREATING' | 'JOB_STATE_RUNNING' | 'JOB_STATE_COMPLETED' | 'JOB_STATE_FAILED' | 'JOB_STATE_CANCELLED' | 'JOB_STATE_DELETING' | 'JOB_STATE_WRITING_RESULTS' | 'JOB_STATE_VALIDATING' | 'JOB_STATE_DELETING_CLEANING_UP' | 'JOB_STATE_PENDING' | 'JOB_STATE_EXPIRED' | 'JOB_STATE_RE_QUEUEING' | 'JOB_STATE_CREATING_INPUT_DATASET' | 'JOB_STATE_IDLE' | 'JOB_STATE_CANCELLING' | 'JOB_STATE_EARLY_STOPPED' | 'JOB_STATE_PAUSED' | 'JOB_STATE_DELETED' | 'JOB_STATE_ARCHIVED'

JobState represents the state an asynchronous job can be in.

  • JOB_STATE_PAUSED: Job is paused, typically due to account suspension or manual intervention.
  • JOB_STATE_DELETED: Job has been deleted.
  • JOB_STATE_ARCHIVED: User-facing state for jobs whose row is retained post-delete (e.g. RLOR trainers within the checkpoint retention window). The internal row is still in JOB_STATE_DELETED; the gateway translates it to ARCHIVED on public responses.
createdBystring

The email address of the user who initiated this fine-tuning job.

evaluatorstring required

The evaluator resource name to use for RLOR fine-tuning job.

outputStatsstring

The output dataset's aggregated stats for the evaluation job.

chunkSizeinteger

Data chunking for rollout, default size 200, enabled when dataset > 300. Valid range is 1-10,000.

outputMetricsstring
maxInferenceReplicaCountinteger
nodeCountinteger

The number of nodes to use for the fine-tuning job. If not specified, the default is 1.

trainerLogsSignedUrlstring

The signed URL for the trainer logs file (stdout/stderr). Only populated if the account has trainer log reading enabled.

acceleratorSecondsobject

Accelerator seconds used by the job, keyed by accelerator type (e.g., "NVIDIA_H100_80GB"). Updated when job completes or is cancelled.

maxConcurrentRolloutsinteger

Maximum number of concurrent rollouts during the RFT job.

maxConcurrentEvaluationsinteger

Maximum number of concurrent evaluations during the RFT job.

purpose'PURPOSE_UNSPECIFIED' | 'PURPOSE_PILOT'

Scheduling purpose for training jobs and deployments.

encryptionState'ENCRYPTION_STATE_UNSPECIFIED' | 'ENCRYPTION_STATE_PLAINTEXT' | 'ENCRYPTION_STATE_CMEK'

EncryptionState is the authoritative, per-resource CMEK marker: the single source of truth for whether a resource's customer-data artifacts are encrypted. It is stamped at resource creation from the owning account's CMEK config and is immutable for the resource's life, so a resource always reads back with the state it was created under regardless of the account's current flag. The encrypt-on-write / decrypt-on-read decision for the resource's data keys off this field.

  • ENCRYPTION_STATE_UNSPECIFIED: Unstamped/legacy rows that predate CMEK. Treated as PLAINTEXT.
  • ENCRYPTION_STATE_PLAINTEXT: Customer-data artifacts are stored as plaintext (Fireworks-owned, no CMEK).
  • ENCRYPTION_STATE_CMEK: Customer-data artifacts are CMEK-encrypted; reads MUST decrypt (fail-closed).