Closed xubenben closed 6 years ago
@xubenben the final evaluation results should be computed against a test dataset with multiple ground truth reference responses. The current code shows not final results since it's only comparing to the test set with 1 reference response.
I recently uploaded the testing dataset with multiple references used in the paper in the data folder. https://github.com/snakeztc/NeuralDialog-CVAE/blob/master/data/test_multi_ref.json.
You can write a script using the same evaluation function to compute the F1 score against this dataset to obtain the final results.
Hi,
I recently running your code, and get this after training with glove, I find it is confusing that the evaluation has no direct connection with results in your paper. Could you give a more detailed explain or more consist evaluation?