Get Training Metrics

Page Outline

• Path Parameters
• Query Parameters
• Behavior
• Response

---

Returns real-time or historical training metrics for a model's latest training run or a specific run, including loss curves and evaluation scores over epochs/iterations.

Base Path: /api/v1/models

Method: GET

Path: /{model_id}/training-metrics

Alias: /{model_id}/training/metrics

Auth: Requires a valid Bearer token (Authorization: Bearer <JWT>)

Path Parameters

Parameter	Type	Description
model_id	string	Unique identifier of the model

Query Parameters

Parameter	Type	Default	Description
run_id	string	—	Specific training run ID. Defaults to the most recent run if omitted
metric	string	—	Filter to a single metric name (e.g., accuracy, loss)
resolution	integer	50	Number of data points to return for time-series metrics (downsampled if run has more)

Behavior

• When a training run is actively in progress, returns live metrics up to the most recent epoch/iteration.
• Metrics are downsampled to resolution points using uniform interval sampling if the raw count exceeds resolution.
• Returns HTTP 404 if the model or requested run ID is not found.
• Alias path /training/metrics is fully supported and returns identical data.

Response

Returns a TrainingMetrics object.

{
  "model_id": "mdl-7f3a19bc",
  "run_id": "run-20260313-a4b9c2",
  "status": "completed",
  "metrics": {
    "accuracy": [0.72, 0.81, 0.88, 0.91, 0.934],
    "loss": [0.65, 0.48, 0.33, 0.22, 0.14],
    "val_accuracy": [0.70, 0.79, 0.86, 0.90, 0.928],
    "val_loss": [0.68, 0.51, 0.37, 0.25, 0.17]
  },
  "epochs_completed": 5,
  "total_epochs": 5
}