Create a tensorflow.data.Dataset from your two questions. Again, labels are not important, so you simply write None $($this is completed for you$),$

use the trained model output to extract v$1,$ v$2($similar to Exercise $04),$

compute the cosine similarity $($dot product$)$ of v$1,$ v$2($similarly to Exercise $04),$

compute res $($the decision if questions are duplicate or not$)$ by comparing d to the threshold.

Replace all "None" instances with correct vaiables!

def predict$($question$1,$ question$2,$ threshold, model, verbose$=$False$)$:

$"""$Function for predicting if two questions are duplicates.

Args:

question$1($str$)$: First question.

question$2($str$)$: Second question.

threshold $($float$)$: Desired threshold.

model $($tensorflow$.$keras.Model$)$: The Siamese model.

data$\_$generator $($function$)$: Data generator function. Defaults to data$\_$generator.

verbose $($bool$,$ optional$)$: If the results should be printed out. Defaults to False.

Returns:

bool: True if the questions are duplicates, False otherwise.

$"""$

generator $=$ tf$.$data.Dataset.from$\_$tensor$\_$slices$((([$question$1],[$question$2]),$None$)).$batch$($batch$\_$size$=1)$

### START CODE HERE ###

# Call the predict method of your model and save the output into v$1$v$2$

v$1$v$2=$ None

# Extract v$1$ and v$2$ from the model output

v$1=$ None

v$2=$ None

# Take the dot product to compute cos similarity of each pair of entries, v$1,$ v$2$

# Since v$1$ and v$2$ are both vectors, use the function tf$.$math.reduce$\_$sum instead of tf$.$linalg.matmul

d $=$ None

# Is d greater than the threshold?

res $=$ None

### END CODE HERE ###

if$($verbose$)$:

print$("$Q$1=",$ question$1,"$

Q$2=",$ question$2)$

print$("$d $=",$ d$.$numpy$())$

print$("$res $=",$ res.numpy$())$

return res.numpy$()$