InvalidArgumentError: required broadcastable shapes [Op:Sub]

[wanderduck]

I keep getting this error: InvalidArgumentError: required broadcastable shapes [Op:Sub] when trying to evaluate a model with a horizon greater than one. Even with the adjusted evaluate_preds() function, this error still occurs.

Can anyone help me figure out a work-around?

[wanderduck]

I had to transpose the data to get it to work/

[mrdbourke]

Glad you got it fixed! Thank you for following up

Shape errors are one of the most common in ML... I run into them every day

[nazzal5448]

i am getting the same error after even transposing the data of model_3_preds and test both or separately

[hemolack]

Can you clarify what you mean by "transpose the data"? I am unable to get the model to fit, even when applying tf.transpose() to either the data or the labels.