Evaluate NeurEco Classification model with GUI#

  • Switch to the Evaluation tab

    **Evaluate** tab for **Tabular Classification**

    Evaluate tab for Tabular Classification#

  • Choose the file to evaluate in Evaluation files section:

    • If the file was supplied in Settings for Build, it is already listed in Evaluation files

    • To add a new file for evaluation, press + in Additional section of Evaluation files

    • For Evaluate, the output file is not required. When it is available, it can be provided for comparison purposes.

  • Once the input file clicked (or a pair input/output), the results of Evaluate are displayed. Here inputs file x_test.csv was added and evaluated:

    Results of **Evaluate** for **Tabular Classification**

    Results of Evaluate for Tabular Classification#

  • To save the results of evaluation into a CSV, NumPy or MAT-file, click Export prediction.
    If the box Also export original files is checked, the input file or input/output files will be exported as well.

Note

By default, the evaluation is performed with the last model available in the checkpoint.
Use the checkpoint slider in the bottom to choose any other available model to Evaluate it.

Note

The results of Evaluate are not one-hot encoded. For each sample its j-th component contains the predicted probability for the sample to belong to the class number j.