The update gate should multiple the previous hidden state, not the candidate state

Course Q&A Deep Learning Specialization Sequence Models
1

In the Full GRU slide (Slide Page 34), I think, the update gate should multiply the previous hidden state, not the candidate state; while (1-updateGate) should multiply the candidate state.

1 Like
2

Thanks for your report.

3

Also, Slide Page 36 also needs to correct