def apply_biases_to_last_layer(graph, counts):
"""
The idea is the following. If the user provides counts file, it is a file that contains log-apriory probabilities,
technically it should be subtracted from the bias of the last layer unless it is a SoftMax.
Case 1:
weights ---\
biases ---\
some layer ---> AffineTransform ---> SoftMax
Then, counts are applied to biases of Affine Transform:
weights ---\
(biases - counts) ---\
some layer ---> AffineTransform ---> SoftMax
Case 2:
weights ---\
biases ---\
some layer ---> AffineTransform
Just takes the last layer and updates biases:
weights ---\
(biases - counts) ---\
some layer ---> AffineTransform
Parameters
----------
graph
counts
Returns
-------
"""""
outputs_ids = find_outputs(graph)
for output in outputs_ids.copy():
node = Node(graph, output)
if node.in_node().op != 'Memory':
continue
outputs_ids.remove(output)
if len(outputs_ids) > 1:
raise Error('Ambiguity in applying counts to several outputs.')
elif len(outputs_ids) == 0:
raise Error('No outputs were found')
node = Node(graph, outputs_ids[0])
target_node = node.in_node()
if target_node and target_node['op'] == 'SoftMax':
data_node = target_node.in_node()
target_node = data_node.in_node()
biases_node = target_node.in_nodes()[2] # first - input, second - weights, third - biases
if biases_node.value is not None:
biases_node.value = np.subtract(biases_node.value, counts) # pylint: disable=assignment-from-no-return
else:
biases_node.value = counts * -1
biases_node.shape = counts.shape
def apply_biases_to_last_layer(graph, counts):
"""
When user provides counts file, it is a file that contains log-apriory probabilities,
technically it should be subtracted from the bias of the last layer unless it is a SoftMax.
Case 1:
weights ---\
biases ---\
some layer ---> AffineTransform ---> SoftMax
Then, counts are applied to biases of Affine Transform:
weights ---\
(biases - counts) ---\
some layer ---> AffineTransform ---> SoftMax
Case 2:
weights ---\
biases ---\
some layer ---> AffineTransform
Just takes the last layer and updates biases:
weights ---\
(biases - counts) ---\
some layer ---> AffineTransform
Parameters
----------
graph
counts
Returns
-------
""" ""
outputs_ids = find_outputs(graph)
for output in outputs_ids.copy():
node = Node(graph, output)
if node.op != 'Memory':
continue
outputs_ids.remove(output)
if len(outputs_ids) > 1:
raise Error('Ambiguity in applying counts to several outputs.')
elif len(outputs_ids) == 0:
raise Error('No outputs were found')
target_node = Node(graph, outputs_ids[0])
if target_node.op == 'SoftMax':
target_node = target_node.in_port(0).get_source().node
sub_node = create_op_node_with_second_input(graph, Add, -counts,
{'name': 'sub_counts'})
target_node.out_port(0).get_connection().set_source(sub_node.out_port(0))
sub_node.in_port(0).connect(target_node.out_port(0))