Move more core code to tf.compat.v1

examples/ImageNetModels/imagenet_utils.py

@@ -13,7 +13,7 @@
 from tensorpack import ModelDesc
 from tensorpack.dataflow import AugmentImageComponent, BatchData, MultiThreadMapData, PrefetchDataZMQ, dataset, imgaug
 from tensorpack.input_source import QueueInput, StagingInput
-from tensorpack.models import regularize_cost
+from tensorpack.models import regularize_cost, l2_regularizer
 from tensorpack.predict import FeedfreePredictor, PredictConfig
 from tensorpack.tfutils.summary import add_moving_summary
 from tensorpack.utils import logger
@@ -339,7 +339,7 @@ def build_graph(self, image, label):
 
         if self.weight_decay > 0:
             wd_loss = regularize_cost(self.weight_decay_pattern,
-                                      tf.contrib.layers.l2_regularizer(self.weight_decay),
+                                      l2_regularizer(self.weight_decay),
                                       name='l2_regularize_loss')
             add_moving_summary(loss, wd_loss)
             total_cost = tf.add_n([loss, wd_loss], name='cost')

examples/basics/mnist-visualizations.py

@@ -98,7 +98,7 @@ def build_graph(self, image, label):
         cost = tf.nn.sparse_softmax_cross_entropy_with_logits(logits=logits, labels=label)
         cost = tf.reduce_mean(cost, name='cross_entropy_loss')
 
-        tf.reduce_mean(tf.cast(tf.nn.in_top_k(logits, label, 1)), tf.float32, name='accuracy')
+        tf.reduce_mean(tf.cast(tf.nn.in_top_k(logits, label, 1), tf.float32), name='accuracy')
 
         wd_cost = tf.multiply(1e-5,
                               regularize_cost('fc.*/W', tf.nn.l2_loss),

tensorpack/graph_builder/training.py

@@ -10,6 +10,7 @@
 import tensorflow as tf
 from six.moves import range, zip
 
+from ..compat import tfv1
 from ..tfutils.common import get_tf_version_tuple
 from ..tfutils.gradproc import ScaleGradient
 from ..tfutils.tower import TrainTowerContext
@@ -101,7 +102,7 @@ def call_for_each_tower(
             device = devices[idx] if devices is not None else '/gpu:{}'.format(t)
             usevs = use_vs[idx] if use_vs is not None else False
             reuse = not usevs and idx > 0
-            with tf.device(device), _maybe_reuse_vs(reuse), TrainTowerContext(
+            with tfv1.device(device), _maybe_reuse_vs(reuse), TrainTowerContext(
                     tower_names[idx],
                     vs_name=tower_names[idx] if usevs else '',
                     index=idx, total=len(towers)):

tensorpack/graph_builder/utils.py

@@ -6,6 +6,7 @@
 from contextlib import contextmanager
 import tensorflow as tf
 
+from ..compat import tfv1
 from ..tfutils.common import get_tf_version_tuple
 from ..tfutils.scope_utils import cached_name_scope, under_name_scope
 from ..tfutils.varreplace import custom_getter_scope
@@ -82,7 +83,7 @@ def __call__(self, op):
         # from tensorflow.python.training.device_util import canonicalize
         # from tensorflow.python.distribute.device_util import canonicalize
         def canonicalize(name):    # tensorflow/tensorflow#11484
-            return tf.DeviceSpec.from_string(name).to_string()
+            return tfv1.DeviceSpec.from_string(name).to_string()
 
         if op.device:
             return op.device

tensorpack/models/nonlin.py

@@ -4,6 +4,7 @@
 
 import tensorflow as tf
 
+from ..compat import tfv1
 from .batch_norm import BatchNorm
 from .common import VariableHolder, layer_register
 
@@ -50,8 +51,8 @@ def PReLU(x, init=0.001, name='output'):
 
     * ``alpha``: learnable slope.
     """
-    init = tf.constant_initializer(init)
-    alpha = tf.get_variable('alpha', [], initializer=init)
+    init = tfv1.constant_initializer(init)
+    alpha = tfv1.get_variable('alpha', [], initializer=init)
     x = ((1 + alpha) * x + (1 - alpha) * tf.abs(x))
     ret = tf.multiply(x, 0.5, name=name)
 

tensorpack/models/regularize.py

@@ -25,7 +25,8 @@ def _log_once(msg):
     l2_regularizer = tf.contrib.layers.l2_regularizer
     l1_regularizer = tf.contrib.layers.l1_regularizer
 else:
-    l2_regularizer = tf.keras.regularizers.l2
+    # oh these little dirty details
+    l2_regularizer = lambda x: tf.keras.regularizers.l2(x * 0.5)  # noqa
     l1_regularizer = tf.keras.regularizers.l1
 
 

tensorpack/predict/concurrency.py

@@ -8,6 +8,7 @@
 import tensorflow as tf
 from six.moves import queue, range
 
+from ..compat import tfv1
 from ..tfutils.model_utils import describe_trainable_vars
 from ..utils import logger
 from ..utils.concurrency import DIE, ShareSessionThread, StoppableThread
@@ -162,7 +163,7 @@ def __init__(self, predictors, batch_size=5):
 
     def start(self):
         if self._need_default_sess:
-            assert tf.get_default_session() is not None, \
+            assert tfv1.get_default_session() is not None, \
                 "Not session is bind to predictors, " \
                 "MultiThreadAsyncPredictor.start() has to be called under a default session!"
         for t in self.threads:

tensorpack/predict/config.py

@@ -3,7 +3,7 @@
 
 
 import six
-import tensorflow as tf
+from ..compat import tfv1 as tf
 
 from ..graph_builder import ModelDescBase
 from ..tfutils import get_default_sess_config

tensorpack/tfutils/argscope.py

@@ -6,7 +6,9 @@
 from contextlib import contextmanager
 from functools import wraps
 from inspect import getmembers, isfunction
+import tensorflow as tf
 
+from ..compat import is_tfv2
 from ..utils import logger
 from .tower import get_current_tower_context
 
@@ -138,6 +140,8 @@ def enable_argscope_for_module(module, log_shape=True):
     Args:
         log_shape (bool): print input/output shapes of each function.
     """
+    if is_tfv2() and module == tf.layers:
+        module = tf.compat.v1.layers
     for name, obj in getmembers(module):
         if isfunction(obj):
             setattr(module, name, enable_argscope_for_function(obj,

tensorpack/tfutils/export.py

@@ -12,6 +12,7 @@
 from tensorflow.python.platform import gfile
 from tensorflow.python.tools import optimize_for_inference_lib
 
+from ..compat import is_tfv2, tfv1
 from ..input_source import PlaceholderInput
 from ..tfutils.common import get_tensors_by_names, get_tf_version_tuple
 from ..tfutils.tower import PredictTowerContext
@@ -60,7 +61,7 @@ def export_compact(self, filename, optimize=True, toco_compatible=False):
 
             self.config.session_init._setup_graph()
             # we cannot use "self.config.session_creator.create_session()" here since it finalizes the graph
-            sess = tf.Session(config=tf.ConfigProto(allow_soft_placement=True))
+            sess = tfv1.Session(config=tfv1.ConfigProto(allow_soft_placement=True))
             self.config.session_init._run_init(sess)
 
             dtypes = [n.dtype for n in input_tensors]
@@ -88,7 +89,7 @@ def export_compact(self, filename, optimize=True, toco_compatible=False):
                 logger.info("Output graph written to {}.".format(filename))
 
     def export_serving(self, filename,
-                       tags=[tf.saved_model.tag_constants.SERVING],
+                       tags=[tf.saved_model.SERVING if is_tfv2() else tf.saved_model.tag_constants.SERVING],
                        signature_name='prediction_pipeline'):
         """
         Converts a checkpoint and graph to a servable for TensorFlow Serving.
@@ -121,21 +122,22 @@ def export_serving(self, filename,
                 self.config.tower_func(*input.get_input_tensors())
 
             input_tensors = get_tensors_by_names(self.config.input_names)
-            inputs_signatures = {t.name: tf.saved_model.utils.build_tensor_info(t) for t in input_tensors}
+            saved_model = tfv1.saved_model.utils
+            inputs_signatures = {t.name: saved_model.build_tensor_info(t) for t in input_tensors}
             output_tensors = get_tensors_by_names(self.config.output_names)
-            outputs_signatures = {t.name: tf.saved_model.utils.build_tensor_info(t) for t in output_tensors}
+            outputs_signatures = {t.name: saved_model.build_tensor_info(t) for t in output_tensors}
 
             self.config.session_init._setup_graph()
             # we cannot use "self.config.session_creator.create_session()" here since it finalizes the graph
-            sess = tf.Session(config=tf.ConfigProto(allow_soft_placement=True))
+            sess = tfv1.Session(config=tfv1.ConfigProto(allow_soft_placement=True))
             self.config.session_init._run_init(sess)
 
-            builder = tf.saved_model.builder.SavedModelBuilder(filename)
+            builder = tfv1.saved_model.builder.SavedModelBuilder(filename)
 
-            prediction_signature = tf.saved_model.signature_def_utils.build_signature_def(
+            prediction_signature = tfv1.saved_model.signature_def_utils.build_signature_def(
                 inputs=inputs_signatures,
                 outputs=outputs_signatures,
-                method_name=tf.saved_model.signature_constants.PREDICT_METHOD_NAME)
+                method_name=tfv1.saved_model.signature_constants.PREDICT_METHOD_NAME)
 
             builder.add_meta_graph_and_variables(
                 sess, tags,

tensorpack/tfutils/gradproc.py

@@ -8,6 +8,7 @@
 import six
 import tensorflow as tf
 
+from ..compat import tfv1
 from ..utils import logger
 from .summary import add_moving_summary
 from .symbolic_functions import print_stat, rms
@@ -40,11 +41,11 @@ def process(self, grads):
 
         # reuse the old name_scope, if process() is called multiple times
         if self._name_scope is None:
-            with tf.name_scope(type(self).__name__) as scope:
+            with tfv1.name_scope(type(self).__name__) as scope:
                 self._name_scope = scope
                 return self._process(grads)
         else:
-            with tf.name_scope(self._name_scope):
+            with tfv1.name_scope(self._name_scope):
                 return self._process(grads)
 
     @abstractmethod
@@ -175,7 +176,7 @@ def _mapper(self, grad, var):
             return grad
         if name not in SummaryGradient._summaried_gradient:
             SummaryGradient._summaried_gradient.add(name)
-            tf.summary.histogram(name + '-grad', grad, collections=self._coll)
+            tfv1.summary.histogram(name + '-grad', grad, collections=self._coll)
             add_moving_summary(rms(grad, name=name + '/rms'))
         return grad
 

tensorpack/tfutils/optimizer.py

@@ -20,7 +20,7 @@ class ProxyOptimizer(tfv1.train.Optimizer):
     A transparent proxy which delegates all methods of :class:`tf.train.Optimizer`
     """
     def __init__(self, opt, name='ProxyOptimizer'):
-        assert isinstance(opt, tf.train.Optimizer), opt
+        assert isinstance(opt, tfv1.train.Optimizer), opt
         super(ProxyOptimizer, self).__init__(False, name)
         self._opt = opt
 

tensorpack/tfutils/sessinit.py

@@ -4,8 +4,8 @@
 import os
 import numpy as np
 import six
-import tensorflow as tf
 
+from ..compat import tfv1 as tf
 from ..utils import logger
 from .common import get_op_tensor_name
 from .varmanip import SessionUpdate, get_checkpoint_path, get_savename_from_varname, is_training_name

tensorpack/tfutils/varmanip.py

@@ -7,6 +7,7 @@
 import six
 import tensorflow as tf
 
+from ..compat import tfv1
 from ..utils import logger
 from .common import get_op_tensor_name
 
@@ -84,7 +85,7 @@ def upcast(vartype, valtype):
             return None
 
         if hasattr(value, 'dtype'):
-            vartype = var.value().dtype
+            vartype = var.dtype
             if vartype != value.dtype:
                 msg = "Variable {} has dtype {} but was given a value of dtype {}.".format(name, vartype, value.dtype)
                 newtype = upcast(var.dtype.base_dtype, value.dtype)
@@ -172,7 +173,7 @@ def get_checkpoint_path(model_path):
     if os.path.basename(model_path) == model_path:
         model_path = os.path.join('.', model_path)  # avoid #4921 and #6142
     if os.path.basename(model_path) == 'checkpoint':
-        assert tf.gfile.Exists(model_path), model_path
+        assert tfv1.gfile.Exists(model_path), model_path
         model_path = tf.train.latest_checkpoint(os.path.dirname(model_path))
         # to be consistent with either v1 or v2
 
@@ -186,7 +187,7 @@ def get_checkpoint_path(model_path):
         logger.info(
             "Checkpoint path {} is auto-corrected to {}.".format(model_path, new_path))
         model_path = new_path
-    assert tf.gfile.Exists(model_path) or tf.gfile.Exists(model_path + '.index'), model_path
+    assert tfv1.gfile.Exists(model_path) or tfv1.gfile.Exists(model_path + '.index'), model_path
     return model_path
 
 
@@ -200,7 +201,7 @@ def load_chkpt_vars(model_path):
         dict: a name:value dict
     """
     model_path = get_checkpoint_path(model_path)
-    reader = tf.train.NewCheckpointReader(model_path)
+    reader = tfv1.train.NewCheckpointReader(model_path)
     var_names = reader.get_variable_to_shape_map().keys()
     result = {}
     for n in var_names:

tox.ini

@@ -10,7 +10,7 @@ exclude = .git,
           examples,
           docs/conf.py
           snippet,
-          examples-old,
+          examples_v2,
           _test.py,
 
 [isort]