FAQs
Installation
Installation | |
---|---|
Does the toolkit run in Jupyter notebooks? | No, the toolkit is a standalone toolkit. Users can use the user interface provided to run the testing workflow. |
Models
Models | |
---|---|
What is the list of models supported by the toolkit? | Currently, the Upload AI Model method supports limited binary classification and regression models trained by algorithms found in scikit-learn, Tensorflow, XGBoost and LightGBM. |
How should I prepare the model to be uploaded? | Visit our How-To guides for preparation of input files: For Tabular/ For Image |
Is there any limit to the model file size? | The maximum size for uploading of the model file and test dataset file is 4 GB each. It is recommended to use a smaller test dataset (i.e., smaller than 4GB) as a large file may take significantly longer to complete. The time taken to run the test will also be affected by the number of features. |
Datasets
Datasets | |
---|---|
Does the toolkit support categorical values in string format? | The toolkit currently does not support columns with categorical values in string format. The toolkit requires these columns to be encoded to numeric values before running any technical tests. |
What pandas version is currently supported by the toolkit? | The toolkit currently runs Pandas v1.5.3. We do not support Pandas 2.x.x. |
Can the toolkit support .csv or .xlsx inputs? | The toolkit supports .csv but not .xlsx. Other Delimiter-separated Values supported are: tab, semicolon, pipe, space, colon |
How should I prepare the dataset to be uploaded? | Visit our How-To guides for preparation of input files: For Tabular/ For Image |
What is considered a good input size for the test data? | Recommended number of test data points is 100, selecting a larger sample will require a longer computation time. |
Results & Reports
Results & Reports | |
---|---|
How do I view test logs after tests have completed running? | Click to edit the project you want to view logs for. In the URL, copy the ID string behind ../project/ Append the ID string to the end of this URL http://localhost:3000/reportStatus/ |
If you have any other questions or need assistance, please check the project discussions or issue tracker for existing threads. If you cannot find a resolution, feel free to create a new discussion or issue, or head over to our contact page if you require assistance.