Start a created fine-tuning job

Prepare training inputs and schedule a workflow for a nonterminal job. running may be set during control-plane claiming and does not establish that GPU training started. Keep querying job_id afterward. Terminal jobs cannot restart.

POST https://api.moi.matrixorigin.cn/v5/workspaces/{id}/finetune-jobs/{job_id}/launch

Before you call this API

Set BASE_URL to the product API prefix https://api.moi.matrixorigin.cn/v5. For private deployments, use the corresponding product API URL.

Prepare identity, training inputs, and environment. Example IDs/results are fixed illustrations. Replace model, run_id, request_id, and compute instance before actual calls;

Request

curl --request POST \
  "$BASE_URL/workspaces/$WORKSPACE_ID/finetune-jobs/$JOB_ID/launch" \
  --header "X-API-Key: $MOI_KEY" \
  --header "X-Workspace-ID: $WORKSPACE_ID"

Parameter

Location

Type

Required

Description

id

Path

string

Yes

Target Workspace ID, matching X-Workspace-ID.

X-Workspace-ID

Headers

string

Yes

Workspace in the same environment accessible to the current identity.

job_id

Path

string

Yes

Fine-tuning job_id returned by successful creation.

Successful response

{
  "code": "OK",
  "msg": "OK",
  "data": {
    "job_id": "ftj-example",
    "workspace_id": "workspace-example",
    "request_id": "request-example",
    "base_model": "example/base-model",
    "status": "running",
    "train_run_ids": [
      "train-run-1",
      "train-run-2"
    ],
    "holdout_run_ids": [
      "holdout-run-1"
    ],
    "recipe": {
      "training_objective": "tool_selector"
    },
    "compute_resource_id": "resource-example",
    "created_at": "2026-11-12T01:00:00Z",
    "updated_at": "2026-11-12T01:00:00Z",
    "mowl_task_id": "workflow-task-example",
    "mowl_case_id": "workflow-case-example",
    "started_at": "2026-11-12T01:00:00Z"
  }
}

Field

Type

Description

code

string

OK on success.

msg

string

Response information.

data

object

Job objects and omission conditions.

Errors and next steps

Lifecycle and recovery distinguishes parameters, identity, conflicts, saved jobs, and actual training results.

Last updated on