Model test results for filling slots with incorrect values

[elnaz776655]

Hi. Thanks for your interesting article. I have a question. In part of the article, you compared a sample conversation from the test dataset with Tripy model and showed that Tripy incorrectly fills slots in the diparture and destination of the taxi domain.
My question is how can we check the results of filling the slots with the wrong value and see them clearly in the conversations?
Do we need to make changes to the code? In which code file can these results be viewed and checked?

[smartyfh]

Hi. Thanks for your interesting article. I have a question. In part of the article, you compared a sample conversation from the test dataset with Tripy model and showed that Tripy incorrectly fills slots in the diparture and destination of the taxi domain.
My question is how can we check the results of filling the slots with the wrong value and see them clearly in the conversations?
Do we need to make changes to the code? In which code file can these results be viewed and checked?

Hi, thanks for your interest! The evaluation.py script will generate full predictions for each slot, then you can compare the predictions with the ground truth. Thanks!

[elnaz776655]

Hi. Thanks for your interesting article. I have a question. In part of the article, you compared a sample conversation from the test dataset with Tripy model and showed that Tripy incorrectly fills slots in the diparture and destination of the taxi domain.
My question is how can we check the results of filling the slots with the wrong value and see them clearly in the conversations?
Do we need to make changes to the code? In which code file can these results be viewed and checked?

Hi, thanks for your interest! The evaluation.py script will generate full predictions for each slot, then you can compare the predictions with the ground truth. Thanks!

Thanks