← All topicsSequence Models
Vanilla RNN
Watch shared weights update hidden memory through time.
A recurrent network carries a hidden state from one observation to the next.
Why do we need this? Sequential predictions need information from earlier observations.
Recurrent connection →Current input / state← Backward sensitivity
STEP 1 OF 20Time 1: Read an input
The sequence supplies one observation at a time.
No account required. Local images and model files stay in your browser. Educational models demonstrate mechanisms; they are not production predictors.