ERROR: No best model saved. Try training for more epochs or add more data to your training set.

Why is my pipeline failing?

Issue Description:

Pipeline run fails, and the following errors are found in chronological order in the Pipeline Report:

  • ERROR: No best model saved. Try training for more epochs or add more data to your training set. NoneType: None
  • ERROR: Training failed for pipeline type: TRAIN_ONLY, error: [Errno 2] No such file or directory: ‘/workspace/model/microservice/models/multi_task_base/network.p’
  • ERROR: Training Job failed, error: [Errno 2] No such file or directory: ‘/workspace/model/microservice/models/multi_task_base/network.p’
  • FileNotFoundError: [Errno 2] No such file or directory: ‘/workspace/model/microservice/models/multi_task_base/network.p’

Troubleshooting:

Search for "Test Set Size" in the Pipeline Report. If one finds the following INFO entry specifying there are zero samples, then the pipeline has failed because the Dataset was exported without a Validation Set:

  • INFO: Test Set Size: 0 samples

Resolution:

  1. Open the Data Labeling Session and select the Training and Validation Set filter or select no filter at all.
  2. Export the Current Search Results.
  3. Create a new Train run Pipeline pointing to the new Dataset.

If you continue to experience issues or have any further questions, please don’t hesitate to contact UiPath Support. Our team is here to assist you and ensure you have the best possible experience with our products.