Rank runs by a metric

Returns runs ranked by a metric aggregation (MIN, MAX, AVG, LAST, VARIANCE) using pre-computed metric summaries. Much faster than comparing individual runs. Useful for finding the best runs in a project by loss, accuracy, or any other metric.

GET /api/runs/leaderboard

Rank runs by a metric

cURL

curl --request GET \
  --url https://pluto-api.trainy.ai/api/runs/leaderboard \
  --header 'Authorization: Bearer <token>'

200

{
  "projectName": "<string>",
  "logName": "<string>",
  "aggregation": "<string>",
  "direction": "<string>",
  "runs": [
    {
      "rank": 123,
      "runId": 123,
      "runName": "<string>",
      "status": "<string>",
      "url": "<string>",
      "value": 123,
      "tags": [
        "<string>"
      ],
      "createdAt": "<string>",
      "config": null
    }
  ],
  "total": 123
}

Authorizations

Authorization · string · header · required

API key obtained from the mlop dashboard

Query Parameters

projectName · string · required

Project name

logName · string · required

Metric name to rank by (e.g., 'train/loss', 'eval/accuracy')

aggregation · enum<string> · default: LAST

Aggregation type: MIN, MAX, AVG, LAST, VARIANCE (default: LAST)

Available options: MIN, MAX, AVG, LAST, VARIANCE

direction · enum<string> · default: ASC

Sort direction: ASC (lowest first) or DESC (highest first). Default: ASC

Available options: ASC, DESC

limit · number | null · default: 20

Number of runs to return (default: 20, max: 100)

offset · number | null · default: 0

Offset for pagination (default: 0)

Response

200 - application/json

Ranked list of runs with metric values

projectName · string · required

logName · string · required

aggregation · string · required

direction · string · required

runs · object[] · required

Show child attributes

total · number · required

Total number of runs that have this metric