Exercise $03$

Instructions: Implement the train$\_$model below to train the neural network above. Here is a list of things you should do:

Compile the model. Here you will need to pass in:

loss$=$TripletLoss

optimizer$=$Adam$()$ with learning rate lr

Call the fit method. You should pass:

train$\_$dataset

epochs

validation$\_$data

You will be using your triplet loss function with Adam optimizer. Also, note that you are not explicitly defining the batch size, because it will be already determined by the Dataset.

This function will return the trained model

# GRADED FUNCTION: train$\_$model

def train$\_$model$($Siamese$,$ TripletLoss, text$\_$vectorizer, train$\_$dataset, val$\_$dataset, d$\_$feature$=128,$ lr$=0\mathrm{.}01,$ train$\_$steps$=5)$:

$"""$Training the Siamese Model

Args:

Siamese $($function$)$: Function that returns the Siamese model.

TripletLoss $($function$)$: Function that defines the TripletLoss loss function.

text$\_$vectorizer: trained instance of $`$TextVecotrization$`$

train$\_$dataset $($tf$.$data.Dataset$)$: Training dataset

val$\_$dataset $($tf$.$data.Dataset$)$: Validation dataset

d$\_$feature $($int$,$ optional$)=$ size of the encoding. Defaults to $128.$

lr $($float$,$ optional$)$: learning rate for optimizer. Defaults to $0\mathrm{.}01$

train$\_$steps $($int$)$: number of epochs

Returns:

tf$.$keras.Model

$"""$

## START CODE HERE ###

# Instantiate your Siamese model

model $=$ Siamese$($None$,$

vocab$\_$size $=$ None, #set vocab$\_$size accordingly to the size of your vocabulary

d$\_$feature $=$ None$)$

# Compile the model

model.compile$($loss$=$None,

optimizer $=$ None

$)$

# Train the model

model.fit$($None$,$

epochs $=$ None,

validation$\_$data $=$ None,

$)$

### END CODE HERE ###

return model