Follow the instruction as documented and run without updating anything, other than the other bug report I created: https://github.com/microsoft/nlp-recipes/issues/614
The experiment fails and from the driver.log it explicitly says it cannot find the training csv file which is apparent from the training script, because those files were never created in earlier step.
Expected behavior (i.e. solution)
Steps to create training/validation dataset should be included. Following the step, the training experiment should succeed.
Description
https://github.com/microsoft/nlp-recipes/blob/master/examples/sentiment_analysis/absa/absa_azureml.ipynb Update data section needs to include how to create training and validation dataset, but the actual code is missing. If you run this notebook as it is, it fails because there are no .csv or .json found as instructed in the training script.
How do we replicate the bug?
Follow the instruction as documented and run without updating anything, other than the other bug report I created: https://github.com/microsoft/nlp-recipes/issues/614 The experiment fails and from the driver.log it explicitly says it cannot find the training csv file which is apparent from the training script, because those files were never created in earlier step.
Expected behavior (i.e. solution)
Steps to create training/validation dataset should be included. Following the step, the training experiment should succeed.
Other Comments