Stat. 654 Handout

Feed-forward Neural Networks and Dropout

```{r}
model <- keras_model_sequential() |>
  layer_dense(units = 512, activation = "relu") |>
  layer_dense(units = 10, activation = "softmax")
```

```{r}
model <- naive_sequential_model(list(
  layer_naive_dense(input_size = 28 * 28, output_size = 512,
                    activation = "op_relu"),
  layer_dropout(rate = 0.4),
  layer_naive_dense(input_size = 512, output_size = 10,
                    activation = "op_softmax"),
))

```