Merge pull request #771 from lukaszkaiser/push

1.6.2

Author: lukaszkaiser

.travis.yml

@@ -21,11 +21,13 @@ matrix:
     - python: "3.6"
       env: TF_VERSION="1.7.*"
 before_install:
-  - echo "deb [arch=amd64] http://storage.googleapis.com/tensorflow-serving-apt stable tensorflow-model-server tensorflow-model-server-universal" | sudo tee /etc/apt/sources.list.d/tensorflow-serving.list
-  - curl https://storage.googleapis.com/tensorflow-serving-apt/tensorflow-serving.release.pub.gpg | sudo apt-key add -
+  # Disabled TensorFlow Serving install until bug fixed. See "Export and query"
+  # section below.
+  # - echo "deb [arch=amd64] http://storage.googleapis.com/tensorflow-serving-apt stable tensorflow-model-server tensorflow-model-server-universal" | sudo tee /etc/apt/sources.list.d/tensorflow-serving.list
+  # - curl https://storage.googleapis.com/tensorflow-serving-apt/tensorflow-serving.release.pub.gpg | sudo apt-key add -
   - sudo apt-get update -qq
   - sudo apt-get install -qq libhdf5-dev
-  - sudo apt-get install -qq tensorflow-model-server
+  # - sudo apt-get install -qq tensorflow-model-server
 install:
   - pip install -q "tensorflow==$TF_VERSION"
   - pip install -q .[tests]

docs/cloud_tpu.md

@@ -18,6 +18,11 @@ See the official tutorial for [running Transformer
 on Cloud TPUs](https://cloud.google.com/tpu/docs/tutorials/transformer)
 for some examples and try out your own problems.
 
+You can train an Automatic Speech Recognition (ASR) model with Transformer
+on TPU by using `transformer` as `model` with `transformer_librispeech_tpu` as
+`hparams_set` and `librispeech` as `problem`. See this [tutorial](tutorials/ast_with_transformer.md) for more details on training it and this
+[notebook](https://colab.research.google.com/github/tensorflow/tensor2tensor/blob/master/tensor2tensor/notebooks/asr_transformer.ipynb) to see how the resulting model transcribes your speech to text.
+
 Image Transformer:
 * `imagetransformer` with `imagetransformer_base_tpu` (or
     `imagetransformer_tiny_tpu`)

docs/index.md

@@ -16,10 +16,14 @@ accessible and [accelerate ML
 research](https://research.googleblog.com/2017/06/accelerating-deep-learning-research.html).
 
 
-## Basics
+## Introduction
 
 * [Walkthrough](walkthrough.md): Install and run.
 * [IPython notebook](https://colab.research.google.com/github/tensorflow/tensor2tensor/blob/master/tensor2tensor/notebooks/hello_t2t.ipynb): Get a hands-on experience.
+* [Automatic Speech Recognition notebook](https://colab.research.google.com/github/tensorflow/tensor2tensor/blob/master/tensor2tensor/notebooks/asr_transformer.ipynb): Transcribe speech to text with a T2T model.
+
+## Basics
+
 * [Overview](overview.md): How all parts of T2T code are connected.
 * [New Problem](new_problem.md): Train T2T models on your data.
 * [New Model](new_model.md): Create your own T2T model.
@@ -29,6 +33,7 @@ research](https://research.googleblog.com/2017/06/accelerating-deep-learning-res
 * [Training on Google Cloud ML](cloud_mlengine.md)
 * [Training on Google Cloud TPUs](cloud_tpu.md)
 * [Distributed Training](distributed_training.md)
+# [Automatic Speech Recognition (ASR) with Transformer](tutorials/asr_with_transformer.md)
 
 ## Solving your task
 

docs/tutorials/asr_with_transformer.md

@@ -1,10 +1,13 @@
 # Automatic Speech Recognition (ASR) with Transformer
 
+Check out the [Automatic Speech Recognition notebook](https://colab.research.google.com/github/tensorflow/tensor2tensor/blob/master/tensor2tensor/notebooks/asr_transformer.ipynb) to see how the resulting model transcribes your speech to text.
+
 ## Data set
 
 This tutorial uses the publicly available
 [Librispeech](http://www.openslr.org/12/) ASR corpus.
 
+
 ## Generate the dataset
 
 To generate the dataset use `t2t-datagen`. You need to create environment

setup.py

@@ -5,7 +5,7 @@
 
 setup(
     name='tensor2tensor',
-    version='1.6.1',
+    version='1.6.2',
     description='Tensor2Tensor',
     author='Google Inc.',
     author_email='[email protected]',

tensor2tensor/data_generators/all_problems.py

@@ -36,6 +36,7 @@
     "tensor2tensor.data_generators.ice_parsing",
     "tensor2tensor.data_generators.imagenet",
     "tensor2tensor.data_generators.imdb",
+    "tensor2tensor.data_generators.lambada",
     "tensor2tensor.data_generators.librispeech",
     "tensor2tensor.data_generators.lm1b",
     "tensor2tensor.data_generators.mnist",

tensor2tensor/data_generators/gym.py

@@ -28,6 +28,7 @@
 from tensor2tensor.data_generators import problem
 from tensor2tensor.data_generators import video_utils
 
+from tensor2tensor.models.research import autoencoders
 from tensor2tensor.models.research import rl
 from tensor2tensor.rl import collect
 from tensor2tensor.rl.envs import tf_atari_wrappers as atari
@@ -42,7 +43,9 @@
 flags = tf.flags
 FLAGS = flags.FLAGS
 
-flags.DEFINE_string("agent_policy_path", "", "File with model for agent")
+
+flags.DEFINE_string("agent_policy_path", "", "File with model for agent.")
+flags.DEFINE_string("autoencoder_path", "", "File with model for autoencoder.")
 
 
 class GymDiscreteProblem(video_utils.VideoProblem):
@@ -179,6 +182,7 @@ class GymPongRandom50k(GymPongRandom5k):
   def num_steps(self):
     return 50000
 
+
 @registry.register_problem
 class GymFreewayRandom5k(GymDiscreteProblem):
   """Freeway game, random actions."""
@@ -209,7 +213,6 @@ def num_steps(self):
     return 50000
 
 
-@registry.register_problem
 class GymDiscreteProblemWithAgent(GymDiscreteProblem):
   """Gym environment with discrete actions and rewards and an agent."""
 
@@ -239,7 +242,7 @@ def _setup(self):
     generator_batch_env = batch_env_factory(
         self.environment_spec, env_hparams, num_agents=1, xvfb=False)
 
-    with tf.variable_scope("", reuse=tf.AUTO_REUSE):
+    with tf.variable_scope(tf.get_variable_scope(), reuse=tf.AUTO_REUSE):
       if FLAGS.agent_policy_path:
         policy_lambda = self.collect_hparams.network
       else:
@@ -252,7 +255,7 @@ def _setup(self):
           create_scope_now_=True,
           unique_name_="network")
 
-    with tf.variable_scope("", reuse=tf.AUTO_REUSE):
+    with tf.variable_scope(tf.get_variable_scope(), reuse=tf.AUTO_REUSE):
       self.collect_hparams.epoch_length = 10
       _, self.collect_trigger_op = collect.define_collect(
           policy_factory, generator_batch_env, self.collect_hparams,
@@ -267,6 +270,22 @@ def restore_networks(self, sess):
           tf.global_variables(".*network_parameters.*"))
       model_saver.restore(sess, FLAGS.agent_policy_path)
 
+  def autoencode(self, image, sess):
+    with tf.Graph().as_default():
+      hparams = autoencoders.autoencoder_discrete_pong()
+      hparams.data_dir = "unused"
+      hparams.problem_hparams = self.get_hparams(hparams)
+      hparams.problem = self
+      model = autoencoders.AutoencoderOrderedDiscrete(
+          hparams, tf.estimator.ModeKeys.EVAL)
+      img = tf.constant(image)
+      img = tf.to_int32(tf.reshape(
+          img, [1, 1, self.frame_height, self.frame_width, self.num_channels]))
+      encoded = model.encode(img)
+      model_saver = tf.train.Saver(tf.global_variables())
+      model_saver.restore(sess, FLAGS.autoencoder_path)
+      return sess.run(encoded)
+
   def generate_encoded_samples(self, data_dir, tmp_dir, unused_dataset_split):
     self._setup()
     self.debug_dump_frames_path = os.path.join(
@@ -275,17 +294,14 @@ def generate_encoded_samples(self, data_dir, tmp_dir, unused_dataset_split):
     with tf.Session() as sess:
       sess.run(tf.global_variables_initializer())
       self.restore_networks(sess)
-      # Actions are shifted by 1 by MemoryWrapper, compensate here.
-      avilable_data_size = sess.run(self.avilable_data_size_op)
-      if avilable_data_size < 1:
-        sess.run(self.collect_trigger_op)
       pieces_generated = 0
-      observ, reward, _, _ = sess.run(self.data_get_op)
       while pieces_generated < self.num_steps + self.warm_up:
         avilable_data_size = sess.run(self.avilable_data_size_op)
         if avilable_data_size < 1:
           sess.run(self.collect_trigger_op)
-        next_observ, next_reward, action, _ = sess.run(self.data_get_op)
+        observ, reward, action, _, img = sess.run(self.data_get_op)
+        if FLAGS.autoencoder_path:
+          observ = self.autoencode(img, sess)
         yield {"image/encoded": [observ],
                "image/format": ["png"],
                "image/height": [self.frame_height],
@@ -294,7 +310,6 @@ def generate_encoded_samples(self, data_dir, tmp_dir, unused_dataset_split):
                "done": [int(False)],
                "reward": [int(reward) - self.min_reward]}
         pieces_generated += 1
-        observ, reward = next_observ, next_reward
 
 
 @registry.register_problem
@@ -318,20 +333,24 @@ def restore_networks(self, sess):
 
 
 @registry.register_problem
-class GymSimulatedDiscreteProblemWithAgentOnPong(GymSimulatedDiscreteProblemWithAgent, GymPongRandom5k):
+class GymSimulatedDiscreteProblemWithAgentOnPong(
+    GymSimulatedDiscreteProblemWithAgent, GymPongRandom5k):
   pass
 
 
 @registry.register_problem
-class GymDiscreteProblemWithAgentOnPong(GymDiscreteProblemWithAgent, GymPongRandom5k):
+class GymDiscreteProblemWithAgentOnPong(
+    GymDiscreteProblemWithAgent, GymPongRandom5k):
   pass
 
 
 @registry.register_problem
-class GymSimulatedDiscreteProblemWithAgentOnFreeway(GymSimulatedDiscreteProblemWithAgent, GymFreewayRandom5k):
+class GymSimulatedDiscreteProblemWithAgentOnFreeway(
+    GymSimulatedDiscreteProblemWithAgent, GymFreewayRandom5k):
   pass
 
 
 @registry.register_problem
-class GymDiscreteProblemWithAgentOnFreeway(GymDiscreteProblemWithAgent, GymFreewayRandom5k):
+class GymDiscreteProblemWithAgentOnFreeway(
+    GymDiscreteProblemWithAgent, GymFreewayRandom5k):
   pass