Merge pull request #449 from rsepassi/push

v1.3

Author: lukaszkaiser

setup.py

@@ -5,7 +5,7 @@
 
 setup(
     name='tensor2tensor',
-    version='1.2.9',
+    version='1.3.0',
     description='Tensor2Tensor',
     author='Google Inc.',
     author_email='[email protected]',

tensor2tensor/bin/t2t-datagen

@@ -112,7 +112,7 @@ _SUPPORTED_PROBLEM_GENERATORS = {
             vocab_filename="vocab.endefr.%d" % 2**15, vocab_size=2**15),
         lambda: audio.timit_generator(
             FLAGS.data_dir, FLAGS.tmp_dir, False, 626,
-            vocab_filename="vocab.endefr.%d" % 2**15, vocab_size=2**15)),     
+            vocab_filename="vocab.endefr.%d" % 2**15, vocab_size=2**15)),
 }
 
 # pylint: enable=g-long-lambda

tensor2tensor/bin/t2t-decoder

@@ -47,8 +47,10 @@ flags = tf.flags
 FLAGS = flags.FLAGS
 
 flags.DEFINE_string("output_dir", "", "Training directory to load from.")
-flags.DEFINE_string("decode_from_file", None, "Path to the source file for decoding")
-flags.DEFINE_string("decode_to_file", None, "Path to the decoded (output) file")
+flags.DEFINE_string("decode_from_file", None,
+                    "Path to the source file for decoding")
+flags.DEFINE_string("decode_to_file", None,
+                    "Path to the decoded (output) file")
 flags.DEFINE_bool("decode_interactive", False,
                   "Interactive local inference mode.")
 flags.DEFINE_integer("decode_shards", 1, "Number of decoding replicas.")

tensor2tensor/bin/t2t-trainer

@@ -61,6 +61,7 @@ flags.DEFINE_string("schedule", "train_and_evaluate",
                     "Method of tf.contrib.learn.Experiment to run.")
 flags.DEFINE_bool("profile", False, "Profile performance?")
 
+
 def main(_):
   tf.logging.set_verbosity(tf.logging.INFO)
   usr_dir.import_usr_dir(FLAGS.t2t_usr_dir)
@@ -85,22 +86,20 @@ def main(_):
   # Run the trainer.
   def run_experiment():
     trainer_utils.run(
-      data_dir=data_dir,
-      model=FLAGS.model,
-      output_dir=output_dir,
-      train_steps=FLAGS.train_steps,
-      eval_steps=FLAGS.eval_steps,
-      schedule=FLAGS.schedule)
-  
+        data_dir=data_dir,
+        model=FLAGS.model,
+        output_dir=output_dir,
+        train_steps=FLAGS.train_steps,
+        eval_steps=FLAGS.eval_steps,
+        schedule=FLAGS.schedule)
+
   if FLAGS.profile:
-    with tf.contrib.tfprof.ProfileContext('t2tprof',
+    with tf.contrib.tfprof.ProfileContext("t2tprof",
                                           trace_steps=range(100),
                                           dump_steps=range(100)) as pctx:
       opts = tf.profiler.ProfileOptionBuilder.time_and_memory()
-      pctx.add_auto_profiling('op', opts, range(100))
-
+      pctx.add_auto_profiling("op", opts, range(100))
       run_experiment()
-
   else:
     run_experiment()
 

tensor2tensor/data_generators/cnn_dailymail.py

@@ -19,10 +19,10 @@
 from __future__ import division
 from __future__ import print_function
 
+import hashlib
 import io
 import os
 import tarfile
-import hashlib
 
 # Dependency imports
 
@@ -129,7 +129,9 @@ def generate_hash(inp):
 
   return filelist
 
+
 def example_generator(all_files, urls_path, sum_token):
+  """Generate examples."""
   def fix_run_on_sents(line):
     if u"@highlight" in line:
       return line
@@ -168,30 +170,37 @@ def fix_run_on_sents(line):
 
     yield " ".join(story) + story_summary_split_token + " ".join(summary)
 
+
 def _story_summary_split(story):
   split_str = u" <summary> "
   split_str_len = len(split_str)
   split_pos = story.find(split_str)
   return story[:split_pos], story[split_pos+split_str_len:]  # story, summary
 
-def write_raw_text_to_files(all_files, urls_path, data_dir, tmp_dir, is_training):
+
+def write_raw_text_to_files(all_files, urls_path, data_dir, tmp_dir,
+                            is_training):
+  """Write text to files."""
   def write_to_file(all_files, urls_path, data_dir, filename):
-    with io.open(os.path.join(data_dir, filename+".source"), "w") as fstory, io.open(os.path.join(data_dir, filename+".target"), "w") as fsummary:
-      for example in example_generator(all_files, urls_path, sum_token=True):
-        story, summary = _story_summary_split(example)
-        fstory.write(story+"\n")
-        fsummary.write(summary+"\n")
+    with io.open(os.path.join(data_dir, filename+".source"), "w") as fstory:
+      with io.open(os.path.join(data_dir, filename+".target"), "w") as fsummary:
+        for example in example_generator(all_files, urls_path, sum_token=True):
+          story, summary = _story_summary_split(example)
+          fstory.write(story+"\n")
+          fsummary.write(summary+"\n")
 
   filename = "cnndm.train" if is_training else "cnndm.dev"
   tf.logging.info("Writing %s" % filename)
   write_to_file(all_files, urls_path, data_dir, filename)
 
   if not is_training:
-    test_urls_path = generator_utils.maybe_download(tmp_dir, "all_test.txt", _TEST_URLS)
+    test_urls_path = generator_utils.maybe_download(
+        tmp_dir, "all_test.txt", _TEST_URLS)
     filename = "cnndm.test"
     tf.logging.info("Writing %s" % filename)
     write_to_file(all_files, test_urls_path, data_dir, filename)
 
+
 @registry.register_problem
 class SummarizeCnnDailymail32k(problem.Text2TextProblem):
   """Summarize CNN and Daily Mail articles to their summary highlights."""
@@ -237,7 +246,8 @@ def generator(self, data_dir, tmp_dir, is_training):
     encoder = generator_utils.get_or_generate_vocab_inner(
         data_dir, self.vocab_file, self.targeted_vocab_size,
         example_generator(all_files, urls_path, sum_token=False))
-    write_raw_text_to_files(all_files, urls_path, data_dir, tmp_dir, is_training)
+    write_raw_text_to_files(all_files, urls_path, data_dir, tmp_dir,
+                            is_training)
     for example in example_generator(all_files, urls_path, sum_token=True):
       story, summary = _story_summary_split(example)
       encoded_summary = encoder.encode(summary) + [EOS]

tensor2tensor/data_generators/generator_utils.py

@@ -447,3 +447,68 @@ def shuffle_dataset(filenames):
     out_fname = fname.replace(UNSHUFFLED_SUFFIX, "")
     write_records(records, out_fname)
     tf.gfile.Remove(fname)
+
+
+def combine_examples_no_inputs(examples, max_length):
+  """Combine examples into longer examples.
+
+  Concatenate targets to form target sequences with length up to max_length.
+  Target sequences longer than max_length are chopped into multiple sequences.
+
+  Args:
+    examples: a generator returning feature dictionaries.
+    max_length: an integer.
+
+  Yields:
+    feature dictionaries.
+  """
+  partial = []
+  for example in examples:
+    x = example["targets"]
+    if len(x) + len(partial) > max_length:
+      if partial:
+        yield {"inputs": [0], "targets": partial}
+        partial = []
+    if len(x) > max_length:
+      num_fragments = len(x) // max_length
+      for i in xrange(num_fragments):
+        yield {"inputs": [0], "targets": x[max_length * i:max_length * (i + 1)]}
+      partial = x[max_length * num_fragments:]
+    else:
+      partial += x
+  if partial:
+    yield {"inputs": [0], "targets": partial}
+
+
+def combine_examples_with_inputs(examples, max_length):
+  """Combine examples into longer examples.
+
+  We combine multiple examples by concatenating the inputs and concatenating
+  the targets.  Sequences where the inputs or the targets are too long are
+  emitted as singletons (not chopped).
+
+  Args:
+    examples: a generator returning feature dictionaries.
+    max_length: an integer.
+
+  Yields:
+    feature dictionaries.
+  """
+  partial_a = []
+  partial_b = []
+  for example in examples:
+    a = example["inputs"]
+    b = example["targets"]
+    if (len(a) + len(partial_a) > max_length or
+        len(b) + len(partial_b) > max_length):
+      if partial_a or partial_b:
+        yield {"inputs": partial_a, "targets": partial_b}
+        partial_a = []
+        partial_b = []
+    if len(a) > max_length or len(b) > max_length:
+      yield {"inputs": a, "targets": b}
+    else:
+      partial_a += a
+      partial_b += b
+  if partial_a or partial_b:
+    yield {"inputs": partial_a, "targets": partial_b}