API Reference

Fine-tuning jobs

Create and manage fine-tuning jobs through the gateway.

POST/v1/fine-tuning/jobs

Submit training and validation data references to create a fine-tuning job.

curl https://<gateway-host>/v1/fine-tuning/jobs \
  -H "Authorization: Bearer $DISRUPTIVERAIN_CLIENT_ID:$DISRUPTIVERAIN_CLIENT_SECRET" \
  -H "Content-Type: application/json" \
  -d '{
    "modelId": "sidis-1",
    "trainingData": {"type": "file", "id": "file_123"},
    "validationData": {"type": "dataset", "id": "dataset_456"},
    "suffix": "sales-tuned",
    "hyperparameters": {
      "epochs": 3,
      "batchSize": 8,
      "learningRate": 0.0001,
      "warmupSteps": 100
    },
    "metadata": {
      "project": "sales-forecast"
    }
  }'

{
  "jobId": "ft_123",
  "modelId": "sidis-1",
  "status": "queued",
  "createdAt": "2024-01-01T00:00:00.000Z",
  "trainingData": {"type": "file", "id": "file_123"},
  "validationData": {"type": "dataset", "id": "dataset_456"},
  "suffix": "sales-tuned",
  "hyperparameters": {
    "epochs": 3,
    "batchSize": 8,
    "learningRate": 0.0001,
    "warmupSteps": 100
  },
  "metadata": {
    "project": "sales-forecast"
  }
}

Successful responses return the job object.

Job statuses

Jobs progress through queued, running, succeeded, failed, or cancelled.

GET/v1/fine-tuning/jobs

Filter by status and paginate with limit and offset.

{
  "jobs": [
    {
      "jobId": "ft_123",
      "modelId": "sidis-1",
      "status": "running",
      "createdAt": "2024-01-01T00:00:00.000Z",
      "trainingData": {"type": "file", "id": "file_123"}
    }
  ],
  "total": 1,
  "limit": 20,
  "offset": 0
}

GET/v1/fine-tuning/jobs/:jobId

Retrieve an individual job by its ID.

POST/v1/fine-tuning/jobs/:jobId/cancel

Cancel a queued or running job. The updated job state is returned in the response body.

Was this page helpful?