[PAD]
[UNK]
[CLS]
[SEP]
[MASK]
[BOS]
[EOS]
the
and
is
are
was
were
be
been
being
have
has
had
do
does
did
will
would
could
should
may
might
can
machine
learning
model
data
train
test
algorithm
neural
network
deep
artificial
intelligence
compute
process
system
method
approach
technique
analysis
result
performance
accuracy
score
metric
evaluation
classification
regression
clustering
similarity
distance
vector
embedding
feature
parameter
weight
bias
layer
hidden
input
output
attention
transformer
bert
roberta
gpt
nlp
text
sentence
word
token
vocabulary
corpus
dataset
training
validation
optimization
gradient
descent
loss
function
objective
minimize
maximize
improve
enhance
better
best
good
bad
high
low
large
small
big
little
new
old
first
last
next
previous
current
recent
future
past
time
number
value
size
length
width
height
depth
dimension
a
an
this
that
these
those
some
many
few
all
each
every
one
two
three
four
five
six
seven
eight
nine
ten
.
,
!
?
-
:
;
'
"
(
)
[
]
{
}
char_161
char_162
char_163