Combine two models at certain layers and add dense layer(s) afterwards.
Usage
merge_models(
models,
layer_names,
layer_dense,
solver = "adam",
learning_rate = 1e-04,
freeze_base_model = c(FALSE, FALSE),
model_seed = NULL
)Arguments
- models
List of two models.
- layer_names
Vector of length 2 with names of layers to merge.
- layer_dense
Vector specifying number of neurons per dense layer after last LSTM or CNN layer (if no LSTM used).
- solver
Optimization method, options are
"adam", "adagrad", "rmsprop"or"sgd".- learning_rate
Learning rate for optimizer.
- freeze_base_model
Boolean vector of length 2. Whether to freeze weights of individual models.
- model_seed
Set seed for model parameters in tensorflow if not
NULL.
Examples
if (FALSE) { # reticulate::py_module_available("tensorflow")
model_1 <- create_model_lstm_cnn(layer_lstm = c(64, 64), maxlen = 50, layer_dense = c(32, 4),
verbose = FALSE)
model_2 <- create_model_lstm_cnn(layer_lstm = c(32), maxlen = 40,
layer_dense = c(8, 2), verbose = FALSE)
# get names of second to last layers
num_layers_1 <- length(model_1$get_config()$layers)
layer_name_1 <- model_1$get_config()$layers[[num_layers_1 - 1]]$name
num_layers_2 <- length(model_2$get_config()$layers)
layer_name_2 <- model_2$get_config()$layers[[num_layers_2 - 1]]$name
# merge models
model <- merge_models(models = list(model_1, model_2),
layer_names = c(layer_name_1, layer_name_2),
layer_dense = c(6, 2),
freeze_base_model = c(FALSE, FALSE))
}