experiments/STT
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Error message if a sample has non-finite loss

Browse Source
This commit is contained in:
Reuben Morais 2019-08-06 14:06:32 +02:00
parent c76070be19
commit 248c01001e
1 changed files with 21 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

29
DeepSpeech.py
View File

@ -214,7 +214,7 @@ def calculate_mean_edit_distance_and_loss(iterator, dropout, reuse):
the decoded result and the batch's original Y.
'''
# Obtain the next batch of data
_, (batch_x, batch_seq_len), batch_y = iterator.get_next()
batch_filenames, (batch_x, batch_seq_len), batch_y = iterator.get_next()
if FLAGS.use_cudnn_rnn:
rnn_impl = rnn_impl_cudnn_rnn
@ -227,11 +227,14 @@ def calculate_mean_edit_distance_and_loss(iterator, dropout, reuse):
# Compute the CTC loss using TensorFlow's `ctc_loss`
total_loss = tfv1.nn.ctc_loss(labels=batch_y, inputs=logits, sequence_length=batch_seq_len)
# Check if any files lead to non finite loss
non_finite_files = tf.gather(batch_filenames, tfv1.where(~tf.math.is_finite(total_loss)))
# Calculate the average loss across the batch
avg_loss = tf.reduce_mean(total_loss)
# Finally we return the average loss
return avg_loss
return avg_loss, non_finite_files
# Adam Optimization
@ -279,6 +282,9 @@ def get_tower_results(iterator, optimizer, dropout_rates):
# Tower gradients to return
tower_gradients = []
# Aggregate any non finite files in the batches
tower_non_finite_files = []
with tfv1.variable_scope(tfv1.get_variable_scope()):
# Loop over available_devices
for i in range(len(Config.available_devices)):
@ -289,7 +295,7 @@ def get_tower_results(iterator, optimizer, dropout_rates):
with tf.name_scope('tower_%d' % i):
# Calculate the avg_loss and mean_edit_distance and retrieve the decoded
# batch along with the original batch's labels (Y) of this tower
avg_loss = calculate_mean_edit_distance_and_loss(iterator, dropout_rates, reuse=i > 0)
avg_loss, non_finite_files = calculate_mean_edit_distance_and_loss(iterator, dropout_rates, reuse=i > 0)
# Allow for variables to be re-used by the next tower
tfv1.get_variable_scope().reuse_variables()
@ -303,13 +309,15 @@ def get_tower_results(iterator, optimizer, dropout_rates):
# Retain tower's gradients
tower_gradients.append(gradients)
tower_non_finite_files.append(non_finite_files)
avg_loss_across_towers = tf.reduce_mean(tower_avg_losses, 0)
tfv1.summary.scalar(name='step_loss', tensor=avg_loss_across_towers, collections=['step_summaries'])
all_non_finite_files = tf.concat(tower_non_finite_files, axis=0)
# Return gradients and the average loss
return tower_gradients, avg_loss_across_towers
return tower_gradients, avg_loss_across_towers, all_non_finite_files
def average_gradients(tower_gradients):
@ -436,7 +444,7 @@ def train():
# Building the graph
optimizer = create_optimizer()
gradients, loss = get_tower_results(iterator, optimizer, dropout_rates)
gradients, loss, non_finite_files = get_tower_results(iterator, optimizer, dropout_rates)
# Average tower gradients across GPUs
avg_tower_gradients = average_gradients(gradients)
@ -517,12 +525,17 @@ def train():
# Batch loop
while True:
try:
_, current_step, batch_loss, step_summary = \
session.run([train_op, global_step, loss, step_summaries_op],
_, current_step, batch_loss, problem_files, step_summary = \
session.run([train_op, global_step, loss, non_finite_files, step_summaries_op],
feed_dict=feed_dict)
except tf.errors.OutOfRangeError:
break
if problem_files.size > 0:
problem_files = [f.decode('utf8') for f in problem_files[..., 0]]
log_error('The following files caused an infinite (or NaN) '
'loss: {}'.format(','.join(problem_files)))
total_loss += batch_loss
step_count += 1