bump version; clean up deprecation

examples/CTC-TIMIT/train-timit.py

@@ -10,7 +10,6 @@ from six.moves import range
 
 from tensorpack import *
 from tensorpack.tfutils.gradproc import GlobalNormClip, SummaryGradient
-from tensorpack.utils import serialize
 
 from timitdata import TIMITBatch
 

examples/FasterRCNN/README.md

@@ -85,7 +85,7 @@ All models are trained with 8 NVIDIA V100s, unless otherwise noted.
 Performance in [Detectron](https://github.com/facebookresearch/Detectron/) can
 be approximately reproduced.
 
- | Backbone                    | mAP<br/>(box;mask)                                             | Detectron mAP <sup>[1](#ft1)</sup><br/> (box;mask) | Time (on 8 V100s) | Configurations <br/> (click to expand)                                                                                                                                                                                                                                                                                                                                                                        |
+ | Backbone                    | mAP<br/>(box;mask)                                             | Detectron mAP <sup>[1](#ft1)</sup><br/> (box;mask) | Time <br/>(on 8 V100s) | Configurations <br/> (click to expand)                                                                                                                                                                                                                                                                                                                                                                        |
  | -                           | -                                                              | -                                                  | -                 | -                                                                                                                                                                                                                                                                                                                                                                                                             |
  | R50-C4                      | 33.5                                                           |                                                    | 17h               | <details><summary>super quick</summary>`MODE_MASK=False FRCNN.BATCH_PER_IM=64`<br/>`PREPROC.TRAIN_SHORT_EDGE_SIZE=600 PREPROC.MAX_SIZE=1024`<br/>`TRAIN.LR_SCHEDULE=[150000,230000,280000]` </details>                                                                                                                                                                                                        |
  | R50-C4                      | 36.6                                                           | 36.5                                               | 44h               | <details><summary>standard</summary>`MODE_MASK=False` </details>                                                                                                                                                                                                                                                                                                                                              |
@@ -98,7 +98,7 @@ be approximately reproduced.
  | R101-C4                     | 41.4;35.2 [:arrow_down:][R101C42x]                             |                                                    | 60h               | <details><summary>standard</summary>`BACKBONE.RESNET_NUM_BLOCKS=[3,4,23,3]` </details>                                                                                                                                                                                                                                                                                                                        |
  | R101-FPN                    | 40.4;36.6 [:arrow_down:][R101FPN2x]                            | 40.9;36.4                                          | 37h               | <details><summary>standard</summary>`MODE_FPN=True`<br/>`BACKBONE.RESNET_NUM_BLOCKS=[3,4,23,3]` </details>                                                                                                                                                                                                                                                                                                    |
  | R101-FPN                    | 46.5;40.1 [:arrow_down:][R101FPN3xCasAug] <sup>[3](#ft3)</sup> |                                                    | 73h               | <details><summary>3x+Cascade+TrainAug</summary>`MODE_FPN=True FPN.CASCADE=True`<br/>`BACKBONE.RESNET_NUM_BLOCKS=[3,4,23,3]`<br/>`TEST.RESULT_SCORE_THRESH=1e-4`<br/>`PREPROC.TRAIN_SHORT_EDGE_SIZE=[640,800]`<br/>`TRAIN.LR_SCHEDULE=[420000,500000,540000]` </details>                                                                                                                                       |
- | R101-FPN<br/>(From Scratch) | 47.5;41.2 [:arrow_down:][R101FPN9xGNCasAugScratch]             | 47.4;40.5<sup>[4](#ft4)</sup>                      | 45h (on 48 V100s) | <details><summary>9x+GN+Cascade+TrainAug</summary>`MODE_FPN=True FPN.CASCADE=True`<br/>`BACKBONE.RESNET_NUM_BLOCKS=[3,4,23,3]`<br/>`FPN.NORM=GN BACKBONE.NORM=GN`<br/>`FPN.FRCNN_HEAD_FUNC=fastrcnn_4conv1fc_gn_head`<br/>`FPN.MRCNN_HEAD_FUNC=maskrcnn_up4conv_gn_head`<br/>`PREPROC.TRAIN_SHORT_EDGE_SIZE=[640,800]`<br/>`TRAIN.LR_SCHEDULE=[1500000,1580000,1620000]`<br/>`BACKBONE.FREEZE_AT=0`</details> |
+ | R101-FPN<br/>(From Scratch) | 47.5;41.2 [:arrow_down:][R101FPN9xGNCasAugScratch]             | 47.4;40.5<sup>[4](#ft4)</sup>                      | 45h <br/>(on 48 V100s) | <details><summary>9x+GN+Cascade+TrainAug</summary>`MODE_FPN=True FPN.CASCADE=True`<br/>`BACKBONE.RESNET_NUM_BLOCKS=[3,4,23,3]`<br/>`FPN.NORM=GN BACKBONE.NORM=GN`<br/>`FPN.FRCNN_HEAD_FUNC=fastrcnn_4conv1fc_gn_head`<br/>`FPN.MRCNN_HEAD_FUNC=maskrcnn_up4conv_gn_head`<br/>`PREPROC.TRAIN_SHORT_EDGE_SIZE=[640,800]`<br/>`TRAIN.LR_SCHEDULE=[1500000,1580000,1620000]`<br/>`BACKBONE.FREEZE_AT=0`</details> |
 
  [R50C42x]: http://models.tensorpack.com/FasterRCNN/COCO-R50C4-MaskRCNN-Standard.npz
  [R50FPN2x]: http://models.tensorpack.com/FasterRCNN/COCO-R50FPN-MaskRCNN-Standard.npz
@@ -119,6 +119,7 @@ For example, our FPN models are sometimes slightly worse in box AP, which is
 
  <a id="ft4">4</a>: This entry does not use ImageNet pre-training. Detectron numbers are taken from Fig. 5 in [Rethinking ImageNet Pre-training](https://arxiv.org/abs/1811.08883).
  Note that our training strategy is slightly different: we enable cascade throughout the entire training.
+ As far as I know, this model is the __best open source model__ on COCO dataset.
 
 ## Notes
 

setup.cfg

-[metadata]
-author = TensorPack contributors
-author-email = ppwwyyxxc@gmail.com
-url = https://github.com/tensorpack/tensorpack
-keywords = tensorflow, deep learning, neural network
-license = Apache
-
-[options]
-zip_safe = False		    # dataset and __init__ use file
-# will call find_packages()
-packages = find:
-
-[wheel]
-universal = 1

setup.py

-import platform
 from os import path
 import setuptools
-from setuptools import setup
+from setuptools import setup, find_packages
 
 version = int(setuptools.__version__.split('.')[0])
 assert version > 30, "Tensorpack installation requires setuptools > 30"
@@ -38,10 +37,20 @@ add_git_version()
 
 setup(
     name='tensorpack',
+    author="TensorPack contributors",
+    author_email="ppwwyyxxc@gmail.com",
+    url="https://github.com/tensorpack/tensorpack",
+    keywords="tensorflow, deep learning, neural network",
+    license="Apache",
+
     version=__version__,   # noqa
-    description='Neural Network Toolbox on TensorFlow',
+    description='A Neural Network Training Interface on TensorFlow',
     long_description=long_description,
     long_description_content_type='text/markdown',
+
+    packages=find_packages(exclude=["examples", "tests"]),
+    zip_safe=False,  		    # dataset and __init__ use file
+
     install_requires=[
         "numpy>=1.14",
         "six",
@@ -61,4 +70,7 @@ setup(
         'all: "linux" in sys_platform': ['python-prctl'],
         'all: python_version < "3.0"': ['tornado'],
     },
+
+    # https://packaging.python.org/guides/distributing-packages-using-setuptools/#universal-wheels
+    options={'bdist_wheel': {'universal': '1'}},
 )

tensorpack/dataflow/common.py

@@ -13,7 +13,6 @@ from six.moves import map, range
 from termcolor import colored
 
 from ..utils import logger
-from ..utils.develop import log_deprecated
 from ..utils.utils import get_rng, get_tqdm, get_tqdm_kwargs
 from .base import DataFlow, DataFlowReentrantGuard, ProxyDataFlow, RNGDataFlow
 
@@ -38,14 +37,10 @@ class TestDataSpeed(ProxyDataFlow):
 
     def __iter__(self):
         """ Will run testing at the beginning, then produce data normally. """
-        self.start_test()
+        self.start()
         for dp in self.ds:
             yield dp
 
-    def start_test(self):
-        log_deprecated("TestDataSpeed.start_test() was renamed to start()", "2019-03-30")
-        self.start()
-
     def start(self):
         """
         Start testing with a progress bar.

tensorpack/dataflow/format.py

@@ -9,7 +9,7 @@ from six.moves import range
 
 from ..utils import logger
 from ..utils.argtools import log_once
-from ..utils.compatible_serialize import loads
+from ..utils.serialize import loads
 from ..utils.develop import create_dummy_class  # noqa
 from ..utils.loadcaffe import get_caffe_pb
 from ..utils.timer import timed_operation

tensorpack/dataflow/serialize.py

@@ -7,7 +7,7 @@ import platform
 from collections import defaultdict
 
 from ..utils import logger
-from ..utils.compatible_serialize import dumps, loads
+from ..utils.serialize import dumps, loads
 from ..utils.develop import create_dummy_class  # noqa
 from ..utils.utils import get_tqdm
 from .base import DataFlow
@@ -150,17 +150,10 @@ class TFRecordSerializer():
             df (DataFlow): the DataFlow to serialize.
             path (str): output tfrecord file.
         """
-        if os.environ.get('TENSORPACK_COMPATIBLE_SERIALIZE', 'msgpack') == 'msgpack':
-            def _dumps(dp):
-                return dumps(dp)
-        else:
-            def _dumps(dp):
-                return dumps(dp).to_pybytes()
-
         size = _reset_df_and_get_size(df)
         with tf.python_io.TFRecordWriter(path) as writer, get_tqdm(total=size) as pbar:
             for dp in df:
-                writer.write(_dumps(dp))
+                writer.write(dumps(dp))
                 pbar.update()
 
     @staticmethod

tensorpack/libinfo.py

@@ -84,5 +84,5 @@ else:
 
 # These lines will be programatically read/write by setup.py
 # Don't touch them.
-__version__ = '0.9.4'
+__version__ = '0.9.5'
 __git_version__ = __version__

tensorpack/models/nonlin.py

@@ -53,7 +53,7 @@ def PReLU(x, init=0.001, name=None):
     * ``alpha``: learnable slope.
     """
     if name is not None:
-        log_deprecated("PReLU(name=...) is deprecated! The output tensor will be named `output`.")
+        log_deprecated("PReLU(name=...)", "The output tensor will be named `output`.")
     init = tfv1.constant_initializer(init)
     alpha = tfv1.get_variable('alpha', [], initializer=init)
     x = ((1 + alpha) * x + (1 - alpha) * tf.abs(x))
@@ -73,7 +73,7 @@ def BNReLU(x, name=None):
         name: deprecated, don't use.
     """
     if name is not None:
-        log_deprecated("BNReLU(name=...) is deprecated! The output tensor will be named `output`.")
+        log_deprecated("BNReLU(name=...)", "The output tensor will be named `output`.")
 
     x = BatchNorm('bn', x)
     x = tf.nn.relu(x, name=name)

tensorpack/models/pool.py

@@ -5,7 +5,6 @@ import numpy as np
 from ..compat import tfv1 as tf  # this should be avoided first in model code
 
 from ..utils.argtools import get_data_format, shape2d
-from ..utils.develop import log_deprecated
 from ._test import TestModel
 from .common import layer_register
 from .shape_utils import StaticDynamicShape
@@ -140,61 +139,6 @@ def FixedUnPooling(x, shape, unpool_mat=None, data_format='channels_last'):
     return ret
 
 
-# Removed (not importable) already; leave it here just for testing purposes.
-def BilinearUpSample(x, shape):
-    """
-    Deterministic bilinearly-upsample the input images.
-    It is implemented by deconvolution with "BilinearFiller" in Caffe.
-    It is aimed to mimic caffe behavior.
-
-    Args:
-        x (tf.Tensor): a NHWC tensor
-        shape (int): the upsample factor
-
-    Returns:
-        tf.Tensor: a NHWC tensor.
-    """
-    log_deprecated("BilinearUpsample", "Please implement it in your own code instead!", "2019-03-01")
-    inp_shape = x.shape.as_list()
-    ch = inp_shape[3]
-    assert ch is not None and ch == 1
-
-    shape = int(shape)
-    filter_shape = 2 * shape
-
-    def bilinear_conv_filler(s):
-        """
-        s: width, height of the conv filter
-        https://github.com/BVLC/caffe/blob/99bd99795dcdf0b1d3086a8d67ab1782a8a08383/include/caffe/filler.hpp#L219-L268
-        """
-        f = np.ceil(float(s) / 2)
-        c = float(2 * f - 1 - f % 2) / (2 * f)
-        ret = np.zeros((s, s), dtype='float32')
-        for x in range(s):
-            for y in range(s):
-                ret[x, y] = (1 - abs(x / f - c)) * (1 - abs(y / f - c))
-        return ret
-    w = bilinear_conv_filler(filter_shape)
-    w = np.repeat(w, ch * ch).reshape((filter_shape, filter_shape, ch, ch))
-
-    weight_var = tf.constant(w, tf.float32,
-                             shape=(filter_shape, filter_shape, ch, ch),
-                             name='bilinear_upsample_filter')
-    x = tf.pad(x, [[0, 0], [shape - 1, shape - 1], [shape - 1, shape - 1], [0, 0]], mode='SYMMETRIC')
-    out_shape = tf.shape(x) * tf.constant([1, shape, shape, 1], tf.int32)
-    deconv = tf.nn.conv2d_transpose(x, weight_var, out_shape,
-                                    [1, shape, shape, 1], 'SAME')
-    edge = shape * (shape - 1)
-    deconv = deconv[:, edge:-edge, edge:-edge, :]
-
-    if inp_shape[1]:
-        inp_shape[1] *= shape
-    if inp_shape[2]:
-        inp_shape[2] *= shape
-    deconv.set_shape(inp_shape)
-    return deconv
-
-
 class TestPool(TestModel):
     def test_FixedUnPooling(self):
         h, w = 3, 4
@@ -213,23 +157,24 @@ class TestPool(TestModel):
         res[0, ::scale, ::scale, :] = 0
         self.assertTrue((res == 0).all())
 
-    def test_BilinearUpSample(self):
-        h, w = 12, 12
-        scale = 2
-
-        mat = np.random.rand(h, w).astype('float32')
-        inp = self.make_variable(mat)
-        inp = tf.reshape(inp, [1, h, w, 1])
-
-        output = BilinearUpSample(inp, scale)
-        res = self.run_variable(output)[0, :, :, 0]
-
-        from skimage.transform import rescale
-        res2 = rescale(mat, scale, mode='edge')
-
-        diff = np.abs(res2 - res)
-
-        # if not diff.max() < 1e-4:
-        #     import IPython
-        #     IPython.embed(config=IPython.terminal.ipapp.load_default_config())
-        self.assertTrue(diff.max() < 1e-4, diff.max())
+# Below was originally for the BilinearUpsample layer used in the HED example
+#     def test_BilinearUpSample(self):
+#         h, w = 12, 12
+#         scale = 2
+#
+#         mat = np.random.rand(h, w).astype('float32')
+#         inp = self.make_variable(mat)
+#         inp = tf.reshape(inp, [1, h, w, 1])
+#
+#         output = BilinearUpSample(inp, scale)
+#         res = self.run_variable(output)[0, :, :, 0]
+#
+#         from skimage.transform import rescale
+#         res2 = rescale(mat, scale, mode='edge')
+#
+#         diff = np.abs(res2 - res)
+#
+#         # if not diff.max() < 1e-4:
+#         #     import IPython
+#         #     IPython.embed(config=IPython.terminal.ipapp.load_default_config())
+#         self.assertTrue(diff.max() < 1e-4, diff.max())

tensorpack/tfutils/symbolic_functions.py

@@ -39,7 +39,7 @@ def rms(x, name=None):
 # don't hurt to leave it here
 @deprecated("Please implement it by yourself.", "2018-04-28")
 def psnr(prediction, ground_truth, maxp=None, name='psnr'):
-    """`Peek Signal to Noise Ratio <https://en.wikipedia.org/wiki/Peak_signal-to-noise_ratio>`_.
+    """`Peak Signal to Noise Ratio <https://en.wikipedia.org/wiki/Peak_signal-to-noise_ratio>`_.
 
     .. math::
 

tensorpack/utils/compatible_serialize.py

-import os
+from .serialize import loads, dumps  # noqa
 
-from .serialize import dumps_msgpack, dumps_pyarrow, loads_msgpack, loads_pyarrow
-
-"""
-Serialization that has compatibility guarantee (therefore is safe to store to disk).
-"""
-
-__all__ = ['loads', 'dumps']
-
-
-# pyarrow has no compatibility guarantee
-# use msgpack for persistent serialization, unless explicitly set from envvar
-if os.environ.get('TENSORPACK_COMPATIBLE_SERIALIZE', 'msgpack') == 'msgpack':
-    loads = loads_msgpack
-    dumps = dumps_msgpack
-else:
-    loads = loads_pyarrow
-    dumps = dumps_pyarrow
+# keep this file for BC