Service Exception 400 InvalidDoctype - Unknown doctype du, not found in model tasks Please train model on given doctype

News Knowledge Base
, , ,
1

Facing issue with ML skill deployment.

Issue Description: Service Exception 400: InvalidDoctype - Unknown doctype du, not found in model tasks. Please train model on given doctype

An error can be observed when the ML Skill created for the same package is failing to deploy with the error “Kubernetes operation failed to create deployment xxxxx_MLSkill_Vxx_xx_xx_xx” and the ML package changed status from “Deployed” to “Deploying”.

Root cause: "Space" in the name of the skills are not allowed - Also no "()" or "(1)" in the skill where we are trying to upload in a different tenant.

A discrepancy in naming conventions between the backend code and the user interface can cause issues when uploading packages. Such naming discrepancies could cause upload failure.

Resolution:

  1. When downloading a package that already exists in the download's folder, the package will automatically be assigned the same name by including (1) in the fine name.
    1. As an example,
  • Version 1 package name: xxxxxxxxx_xxxxx_MLPACKAGE_Vxx_xx_xx (1).zip
    • Version 4 package name: xxxxxxxxx_xxxxx_MLPACKAGE_Vxx_xx_xx.zip
      • The space and parentheses in Version 1 are causing the download issue: MLPACKAGE_V23_4_1 (1).zip

So, because of the space, the curl command gets confused and outputs that shady "unary operator expected".

Contact UiPath Product Support: If the issue persists, create UiPath Support Ticket for further troubleshooting.

Ensure to reproduce the issue and provide/attach the below list of details when opening a ticket:

  1. Was the re-deployment of the same package attempted? If not, it is recommended.
  2. Share a screenshot of the ML package page.
  3. Provide details of the base model used for training the pipeline (For example: ML Packages/Out of the box Packages/UiPath Document Understanding/Invoices version 22.10.1.0).
  4. Attach the ML Skill Screenshot and the ML Skill report.
  5. If cloud AI Center is used, provide the below details:
  • Navigate to Admin -> Organization Settings -> Copy the "URL" and "Support ID" -> Share them with us.
    • Open AIC, click on three dots near the tenant selection dropdown, click on view profile, copy the account ID and tenant ID and attach them to the ticket.
    • Share the project name as well.