User Guides

Troubleshooting

Resolve common user-facing errors quickly.

Quickstart
Get from sign-in to a completed first job.
Choosing a Service
Match your data and goal to the right workflow.
Data Preparation
Prepare tabular datasets and media manifests for successful uploads.
Jobs and Results
Understand statuses, reports, SHAP, and outputs.
Billing and Credits
How charges, estimates, and refunds work.
Prediction Guide
Use trained jobs for manual and batch prediction.
Troubleshooting
Resolve common user-facing errors quickly.
FAQ
Short answers to the most common platform questions.
Reference
Handbook and Technical Reference

Open the handbook for the full product guide, then use the model-method appendix when you need mathematical details.

Open Handbook Open model-method appendix

Troubleshooting

Upload issues

  • Check required columns
  • Confirm file type
  • Remove malformed rows

If the upload page keeps rejecting the file: - compare the file against the service template - confirm the selected service matches the data shape - simplify the file before debugging advanced settings

Training issues

  • Re-run with the balanced comparison strategy
  • Reduce advanced overrides
  • Confirm the selected service matches the dataset shape

If training completes but the result looks wrong: - confirm the target column selected in the form - review whether identifier or leakage columns were included - confirm the best-model metric family matches the service type

SHAP issues

  • Some services run SHAP only on demand
  • Some model types do not support SHAP cleanly
  • A training success does not guarantee SHAP success

If SHAP is missing: - check whether SHAP is an add-on for that service - confirm the selected model supports the explainability method - review the job page for SHAP status rather than assuming training included it

Manuscript issues

  • AI Report/manuscript generation requires an explicit user request from the completed job page/API
  • A fallback manuscript can still be structurally correct but weaker in prose quality
  • If the manuscript shows the wrong metric family, the job may have been generated before a metric-mapping fix or through the wrong service path

If a manuscript looks wrong: 1. confirm the service type on the job 2. confirm the saved objective 3. confirm SHAP or feature-importance artifacts are present when expected 4. regenerate after the underlying job artifacts are complete

Prediction issues

  • Confirm the trained model artifact exists
  • Confirm preprocessing files exist
  • Use the exact expected schema for batch prediction

Billing issues

  • Compare job detail billing with the account Billing page
  • Distinguish estimates from recorded debits
  • Check whether a failed add-on action was refunded

Session issues

  • If your session expires, sign in again and return to the same job page

When to escalate

Escalate or investigate deeper when: - the job service type and displayed metric family disagree - a completed job is missing core artifacts - billing records do not line up with triggered actions - repeated manuscript generations keep using fallback because provider access is failing

Good debugging order

  1. Verify service type
  2. Verify target and required columns
  3. Verify training artifacts
  4. Verify SHAP status
  5. Verify billing events
  6. Verify manuscript generation provenance

Related guides

  • Billing and Credits
  • Jobs and Results
  • Prediction Guide
  • FAQ