clean-up some deprecations

examples/GAN/README.md

@@ -18,7 +18,7 @@ Reproduce the following GAN-related methods:
 
 + BEGAN ([BEGAN: Boundary Equilibrium Generative Adversarial Networks](https://arxiv.org/abs/1703.10717))
 
-Please see the __docstring__ in each script for detailed usage and pretrained models.
+Please see the __docstring__ in each script for detailed usage and pretrained models. MultiGPU training is supported.
 
 ## DCGAN.py
 

tensorpack/models/model_desc.py

@@ -10,14 +10,11 @@ import six
 
 from ..utils import logger
 from ..utils.naming import INPUTS_KEY
-from ..utils.develop import deprecated, log_deprecated
 from ..utils.argtools import memoized
 from ..tfutils.model_utils import apply_slim_collections
 
 __all__ = ['InputDesc', 'InputVar', 'ModelDesc', 'ModelFromMetaGraph']
 
-# TODO "variable" is not the right name to use for input here.
-
 
 class InputDesc(object):
     """ Store metadata about input placeholders. """
@@ -50,7 +47,8 @@ class InputVar(InputDesc):
 
 @six.add_metaclass(ABCMeta)
 class ModelDesc(object):
-    """ Base class for a model description """
+    """ Base class for a model description.
+    """
 
 # inputs:
     @memoized
@@ -63,11 +61,6 @@ class ModelDesc(object):
         """
         return self.build_placeholders()
 
-    @deprecated("Use get_reused_placehdrs() instead.", "2017-04-11")
-    def get_input_vars(self):
-        # this wasn't a public API anyway
-        return self.get_reused_placehdrs()
-
     def build_placeholders(self, prefix=''):
         """
         For each InputDesc, create new placeholders with optional prefix and
@@ -76,12 +69,12 @@ class ModelDesc(object):
         Returns:
             list[tf.Tensor]: the list of built placeholders.
         """
-        input_vars = self._get_inputs()
-        for v in input_vars:
+        inputs = self._get_inputs()
+        for v in inputs:
             tf.add_to_collection(INPUTS_KEY, v.dumps())
         ret = []
         with tf.name_scope(None):   # clear any name scope it might get called in
-            for v in input_vars:
+            for v in inputs:
                 placehdr_f = tf.placeholder if not v.sparse else tf.sparse_placeholder
                 ret.append(placehdr_f(
                     v.type, shape=v.shape,
@@ -95,15 +88,11 @@ class ModelDesc(object):
         """
         return self._get_inputs()
 
-    def _get_inputs(self):  # this is a better name than _get_input_vars
+    @abstractmethod
+    def _get_inputs(self):
         """
         :returns: a list of InputDesc
         """
-        log_deprecated("", "_get_input_vars() was renamed to _get_inputs().", "2017-04-11")
-        return self._get_input_vars()
-
-    def _get_input_vars(self):  # keep backward compatibility
-        raise NotImplementedError()
 
     def build_graph(self, model_inputs):
         """
@@ -142,8 +131,8 @@ class ModelDesc(object):
     def get_optimizer(self):
         """
         Return the optimizer used in the task.
-        Used by some of the tensorpack :class:`Trainer` which only uses a single optimizer.
-        You can ignore this method if you use your own trainer with more than one optimizers.
+        Used by some of the tensorpack :class:`Trainer` which assume single optimizer.
+        You can (and should) ignore this method if you use a custom trainer with more than one optimizers.
 
         Users of :class:`ModelDesc` will need to implement `_get_optimizer()`,
         which will only be called once per each model.
@@ -157,6 +146,9 @@ class ModelDesc(object):
         raise NotImplementedError()
 
     def get_gradient_processor(self):
+        return self._get_gradient_processor()
+
+    def _get_gradient_processor(self):
         return []
 
 

tensorpack/train/config.py

@@ -2,8 +2,6 @@
 # File: config.py
 # Author: Yuxin Wu <ppwwyyxxc@gmail.com>
 
-import tensorflow as tf
-
 from ..callbacks import (
     Callbacks, MovingAverageSummary,
     ProgressBar, MergeAllSummaries,
@@ -15,7 +13,6 @@ from ..utils.develop import log_deprecated
 from ..tfutils import (JustCurrentSession,
                        get_default_sess_config, SessionInit)
 from ..tfutils.sesscreate import NewSessionCreator
-from ..tfutils.optimizer import apply_grad_processors
 from .input_source import InputSource
 
 __all__ = ['TrainConfig']
@@ -154,15 +151,9 @@ class TrainConfig(object):
         assert len(set(self.predict_tower)) == len(self.predict_tower), \
             "Cannot have duplicated predict_tower!"
 
-        if 'optimizer' in kwargs:
-            log_deprecated("TrainConfig(optimizer=...)",
-                           "Use ModelDesc._get_optimizer() instead.",
-                           "2017-04-12")
-            self._optimizer = kwargs.pop('optimizer')
-            assert_type(self._optimizer, tf.train.Optimizer)
-        else:
-            self._optimizer = None
-
+        assert 'optimizer' not in kwargs, \
+            "TrainConfig(optimizer=...) was already deprecated! " \
+            "Use ModelDesc._get_optimizer() instead."
         assert len(kwargs) == 0, 'Unknown arguments: {}'.format(str(kwargs.keys()))
 
     @property
@@ -176,19 +167,3 @@ class TrainConfig(object):
     @property
     def callbacks(self):        # disable setter
         return self._callbacks
-
-    @property
-    def optimizer(self):
-        """ for back-compatibilty only. will remove in the future"""
-        if self._optimizer:
-            opt = self._optimizer
-        else:
-            opt = self.model.get_optimizer()
-        gradproc = self.model.get_gradient_processor()
-        if gradproc:
-            log_deprecated("ModelDesc.get_gradient_processor()",
-                           "Use gradient processor to build an optimizer instead.", "2017-04-12")
-            opt = apply_grad_processors(opt, gradproc)
-        if not self._optimizer:
-            self._optimizer = opt
-        return opt

tensorpack/train/feedfree.py

@@ -64,7 +64,7 @@ class SingleCostFeedfreeTrainer(FeedfreeTrainerBase):
         """ get the cost and gradient"""
         self.build_train_tower()
         cost = self.model.get_cost()    # assume single cost
-        opt = self.config.optimizer     # TODO XXX
+        opt = self.model.get_optimizer()
         # GATE_NONE faster?
         grads = opt.compute_gradients(
             cost,
@@ -96,7 +96,8 @@ class SimpleFeedfreeTrainer(SingleCostFeedfreeTrainer):
         super(SimpleFeedfreeTrainer, self)._setup()
         with TowerContext('', is_training=True):
             cost, grads = self._get_cost_and_grad()
-        self.train_op = self.config.optimizer.apply_gradients(grads, name='min_op')
+        opt = self.model.get_optimizer()
+        self.train_op = opt.apply_gradients(grads, name='min_op')
         # skip training
         # self.train_op = tf.group(*self._input_tensors)
 

tensorpack/train/input_source.py

@@ -364,8 +364,8 @@ class StagingInputWrapper(FeedfreeInput):
             devices: list of devices to be used for each training tower
             nr_stage: number of elements to prefetch
         """
+        assert isinstance(input, FeedfreeInput), input
         self._input = input
-        assert isinstance(input, FeedfreeInput)
         self._devices = devices
         self._nr_stage = nr_stage
         self._areas = []

tensorpack/train/multigpu.py

@@ -167,7 +167,7 @@ class SyncMultiGPUTrainerParameterServer(MultiGPUTrainerBase, SingleCostFeedfree
         grads = SyncMultiGPUTrainerParameterServer._average_grads(grad_list)
         # grads = grad_list[0]
 
-        self.train_op = self.config.optimizer.apply_gradients(grads, name='min_op')
+        self.train_op = self.model.get_optimizer().apply_gradients(grads, name='min_op')
 
 
 def SyncMultiGPUTrainer(config):
@@ -217,7 +217,8 @@ class AsyncMultiGPUTrainer(MultiGPUTrainerBase,
             grad_list = [gradproc.process(gv) for gv in grad_list]
 
         # use grad from the first tower for iteration in main thread
-        self.train_op = self.config.optimizer.apply_gradients(grad_list[0], name='min_op')
+        self._opt = self.model.get_optimizer()
+        self.train_op = self._opt.apply_gradients(grad_list[0], name='min_op')
 
         self._start_async_threads(grad_list)
 
@@ -227,7 +228,7 @@ class AsyncMultiGPUTrainer(MultiGPUTrainerBase,
         self.async_step_counter = itertools.count()
         self.training_threads = []
         for k in range(1, self.config.nr_tower):
-            train_op = self.config.optimizer.apply_gradients(grad_list[k])
+            train_op = self._opt.apply_gradients(grad_list[k])
 
             def f(op=train_op):  # avoid late-binding
                 self.sess.run([op])         # TODO this won't work with StageInput

tensorpack/train/trainer.py

@@ -42,5 +42,5 @@ class SimpleTrainer(Trainer):
             model.build_graph(self.inputs)
             cost_var = model.get_cost()
 
-        opt = self.config.optimizer
+        opt = self.model.get_optimizer()
         self.train_op = opt.minimize(cost_var, name='min_op')

tests/dev/git-hooks/pre-commit

@@ -4,7 +4,9 @@ flake8 .
 cd examples
 GIT_ARG="--git-dir ../.git --work-tree .."
 
-# find out modified python files
+# find out modified python files, so that we ignored unstaged files
 MOD=$(git $GIT_ARG status -s | grep -E '\.py$'  | grep -E '^ *M|^ *A ' | cut -c 4-)
 # git $GIT_ARG status -s | grep -E '\.py$'
-flake8 $MOD
+if [[ -n $MOD ]]; then
+	flake8 $MOD
+fi