python - Tensorflow onehot encode -

i'm new @ using tensorflow , have questions tensorflows 1 hot encoding.

i want read cvs file in in last column represents labels. labels integer values 1 7.

i want classification using softmax model.

therefore need labels in onehot tensor format?

is there simple way/tensorflow built-in convert labels 1 hot?

according tutorial file parsing have following code parsing csv part of 1 hot encoding missing.

def read_from_cvs(filename_queue):     reader = tf.textlinereader()     key, value = reader.read(filename_queue)     record_defaults = [[] col in range((num_attributes))] # no defaults, values must given     attributes = tf.decode_csv(value, record_defaults=record_defaults)     features = tf.stack(attributes[1:-1])     labels = tf.stack(attributes[-1])     return features, labels  def input_pipeline(filename = 'dataset.csv', batch_size = 30, num_epochs=none):     filename_queue = tf.train.string_input_producer(filename, num_epochs=num_epochs, shuffle=true)     features, labels = read_from_cvs(filename_queue)      min_after_dequeue = 10000     capacity = min_after_dequeue + 3 * batch_size     feature_batch, label_batch = tf.train.shuffle_batch(         [features, labels], batch_size=batch_size, capacity=capacity,         min_after_dequeue=min_after_dequeue)     return feature_batch, label_batch 

you can use tf.one_hot

# depth = num_clasess in general problem      labels = tf.one_hot(tf.stack(attributes[-1]), depth)