Experiments
List experiments
Changed onget/api/v1/experiments
Request
- Base URL: / (relative, and the document was published on a file host, not at the API’s origin)
- URL: /api/v1/experiments
- Auth: none declared
Query parameters
ownerstring
Admin-only owner filter.
qstring
Fuzzy keyword.
limitinteger
Page size (default 50, max 200).
continuestring
Opaque continue token from the previous response.
Headers
X-Axisml-Tenantstring
Active tenant for the request (fallback for the axisml.tenant cookie).
Response
A page of experiments.
Example response
{
"continueToken": "",
"count": 1,
"items": [
{
"annotations": {
"axisml.io/created-by": "li.wei"
},
"createdAt": "2026-06-20T08:00:00Z",
"description": "Training experiment fine-tuning BERT on a Chinese corpus.",
"displayName": "BERT fine-tuning experiment",
"id": "d4f8a1b2-3c5e-4a7b-9c0d-1e2f3a4b5c6d",
"labels": {
"team": "nlp"
},
"name": "bert-finetune",
"namespace": "team-nlp",
"owner": "zhang.san",
"ownerId": "3a2b1c0d-4e5f-6789-abcd-ef0123456789",
"runSummary": {
"active": 1,
"count": 5,
"latestPhase": "Succeeded",
"latestRunAt": "2026-06-28T09:25:00Z",
"recent": [
"Succeeded",
"Running",
"Succeeded",
"Failed",
"Succeeded"
]
},
"spec": {
"artifacts": [
{
"kind": "model",
"name": "bert-base",
"version": "2.1.0"
}
],
"backend": {
"engine": "pytorchjob",
"name": "kubeflow-trainer"
},
"poolName": "gpu-a100",
"roles": [
{
"name": "worker",
"replicas": 2,
"restartPolicy": "OnFailure",
"template": {
"args": [
"--lr",
"0.001",
"--epochs",
"10",
"--batch-size",
"64"
],
"command": [
"python",
"train.py"
],
"env": [
{
"name": "WANDB_MODE",
"value": "offline"
}
],
"image": "registry.axisml.io/training/bert:2.1.0",
"resources": {
"cpu": "8",
"memory": "64Gi",
"nvidia.com/gpu": "2"
}
}
}
],
"runPolicy": {
"activeDeadlineSeconds": 86400,
"backoffLimit": 2,
"ttlSecondsAfterFinished": 3600
},
"unitName": "a100-2x"
},
"tenantName": "team-nlp",
"updatedAt": "2026-06-28T09:30:00Z"
}
],
"partial": false
}Changes
- warning
added the new
Queuedenum value to the///////////response property for the response status - warning
added the new
Queuedenum value to the////////response property for the response status
- warning
- info
added the optional property
///////to the response with the status - info
added the optional property
//////////////to the response with the status
- info