Getting this InvalidArgumentError: “required broadcatable shapes”. After some debugging I got that the y_true shape and y_pred_class shape is not same in the test set “w2_unittest.test_masked_accuracy”. Here, nothing can be done by me, it’s I think mistake of problem-setter. How to solve this?
How are they different? This might indicate where you got it wrong. Maybe when predicting, you set the axis= argument wrong? (the argmax should be on the last axis).
Cheers
1 Like