Tying weights and shape debugging

[junpenglao]

I usually check the shape by doing w1_repeat.tag.test_value.shape

The observed of a Categorical distribution should be a vector, replace the one hot encoded Y with np.where(Y==1)[1] should work.