AutoTrain training stopped/failed without explanation

🤗AutoTrain
1

I created an AutoTrain project, started a training run, and I get this:

image
image1750Ă—692 94.2 KB

Nothing here is clickable or provides any other information. The JS console says nothing.

What does “stopped” mean on a training run? What does “failed” mean on a project? How do I get past this error?

1 Like
2

I then started a new AutoTrain project, then uploaded the same training set in a different format (csv rather than jsonl). But I get the same error after starting the training run.

3

This is all that’s in the response to https://api.autotrain.huggingface.co/projects/69004:

{
    "id": 69004,
    "username": "jameshfisher",
    "proj_name": "t3",
    "task": 2,
    "status": 9,
    "training_status": "failed",
    "last_data_step": {
        "step_name": "save_metadata",
        "status": "success",
        "started_at": "2023-06-22T17:03:08.731553",
        "finished_at": "2023-06-22T17:03:08.743766",
        "error": null
    },
    "created_at": "2023-06-22T17:02:46.850487",
    "updated_at": "2023-06-22T17:03:36.231645",
    "dataset_id": "jameshfisher/autotrain-data-t3",
    "config": {
        "language": "en",
        "advanced": false,
        "autotrain": true,
        "max_models": 1,
        "hub_model": null,
        "instance": {
            "provider": "ovh",
            "instance_type": "p3",
            "max_runtime_seconds": 172800,
            "num_instances": 1,
            "disk_size_gb": 150
        },
        "evaluation": {
            "models": [],
            "metrics": [],
            "hf_username": null
        },
        "benchmark": {
            "model": "",
            "dataset": "",
            "submission_dataset": "",
            "submission_id": "",
            "create_prediction_repo": true
        },
        "params": []
    }
}
2 Likes
4

I am experiencing the same issue.

1 Like
5

I’m in the same boat.

Is this problem related to the data we have put in ?