I am trying to train the AI fabric EnglishTextClassification model with the Train Pipeline. But after sometime it kept on failing saying that “No valid data to run this pipeline.”(error details attached). I’m uploading a csv as dataset(attached it). I’m not exacly sure of the file format, but I’ve given it according to a video in youtube. Please look into it and revertfiles.zip (1.7 KB) . Also let me know where can I found these details like data format etc., as I searched for it but didn’t found anything.
If you look at the full logs (go to pipeline page, click three dots button on the grid, details, download full logs) you will see this line in your file:
2020-07-30 13:10:39,324 - aiflib.data_manager:info:15 - INFO: Provided data does not have enough training examples to train, you must provide at least 5 training examples of each class
On your dataset each class has only 3 data points.
Full documentation of model is there; https://docs.uipath.com/ai-fabric/v0/docs/english-text-classification although I see that this requirement is missing, we’ll add it.
Thanks Jeremy. With this log message it was easy to figure out about the issue. But I didn’t expect to check for the logs in another place apart from ML logs. Anyway, thanks for sharing the documentation.
Can you please let me know whats the solution.I am getting same sort of error.
Please refer this post for details
Can you put target column first and then input. That might be the reason. Also check the log file location specified by Jeremy to get more details about the error, if changing the sequence of the columns does not help.