[Relay] Conv2D padding representation (apache#4787)

* enforce 4-way padding

* add util with get_pad_tuple

* delete unnecessary arguments

* fix lint

* add container.Array case

* fix cudnn conv2d asymmetric padding logic

* rename get_pad_tuple to get_pad_tuple2d

* revert change for topi/python/topi/nn/conv2d.py

* add get_pad_tuple2d for several contrib conv2d ops

* add get_pad_tuple2d for all conv2d ops

python/tvm/relay/op/nn/nn.py

@@ -19,6 +19,7 @@
 from __future__ import absolute_import as _abs
 from ...expr import TupleWrapper
 from . import _make
+from .util import get_pad_tuple2d
 
 
 def conv1d(data,
@@ -200,8 +201,9 @@ def conv2d(data,
         strides = (strides, strides)
     if isinstance(dilation, int):
         dilation = (dilation, dilation)
-    if isinstance(padding, int):
-        padding = (padding, padding)
+    # TODO enforce 4-way padding in topi/nn/conv2d after #4644 merged
+    # convert 2-way padding to 4-way padding
+    padding = get_pad_tuple2d(padding)
 
     return _make.conv2d(data, weight, strides, padding, dilation,
                         groups, channels, kernel_size, data_layout,
@@ -363,6 +365,8 @@ def conv2d_transpose(data,
     result : tvm.relay.Expr
         The computed result.
     """
+    # convert 2-way padding to 4-way padding
+    padding = get_pad_tuple2d(padding)
     return _make.conv2d_transpose(data, weight, strides, padding, dilation,
                                   groups, channels, kernel_size, data_layout,
                                   kernel_layout, out_layout, output_padding, out_dtype)
@@ -1758,6 +1762,8 @@ def contrib_conv2d_winograd_without_weight_transform(data,
     result : tvm.relay.Expr
         The computed result.
     """
+    # convert 2-way padding to 4-way padding
+    padding = get_pad_tuple2d(padding)
     return _make.contrib_conv2d_winograd_without_weight_transform(
         data, weight, tile_size, strides, padding, dilation,
         groups, channels, kernel_size, data_layout,
@@ -1824,6 +1830,8 @@ def contrib_conv2d_winograd_nnpack_without_weight_transform(data,
     result : tvm.relay.Expr
         The computed result.
     """
+    # convert 2-way padding to 4-way padding
+    padding = get_pad_tuple2d(padding)
     return _make.contrib_conv2d_winograd_nnpack_without_weight_transform(
         data, weight, strides, padding, dilation,
         groups, channels, kernel_size, data_layout,
@@ -1891,6 +1899,8 @@ def contrib_conv2d_nchwc(data,
     result : tvm.relay.Expr
         The computed result.
     """
+    # convert 2-way padding to 4-way padding
+    padding = get_pad_tuple2d(padding)
     return _make.contrib_conv2d_NCHWc(data, kernel, strides, padding, dilation,
                                       groups, channels, kernel_size, data_layout,
                                       kernel_layout, out_layout, out_dtype)
@@ -1956,6 +1966,8 @@ def contrib_depthwise_conv2d_nchwc(data,
     result : tvm.relay.Expr
         The computed result.
     """
+    # convert 2-way padding to 4-way padding
+    padding = get_pad_tuple2d(padding)
     return _make.contrib_depthwise_conv2d_NCHWc(data, kernel, strides, padding, dilation,
                                                 groups, channels, kernel_size, data_layout,
                                                 kernel_layout, out_layout, out_dtype)
@@ -2021,6 +2033,8 @@ def contrib_conv2d_nchwc_int8(data,
     result : tvm.relay.Expr
         The computed result.
     """
+    # convert 2-way padding to 4-way padding
+    padding = get_pad_tuple2d(padding)
     return _make.contrib_conv2d_NCHWc_int8(data, kernel, strides, padding, dilation,
                                            groups, channels, kernel_size, data_layout,
                                            kernel_layout, out_layout, out_dtype)
@@ -2142,6 +2156,8 @@ def deformable_conv2d(data,
         The computed result.
 
     """
+    # convert 2-way padding to 4-way padding
+    padding = get_pad_tuple2d(padding)
     return _make.deformable_conv2d(data, offset, weight, strides, padding, dilation,
                                    deformable_groups, groups, channels, kernel_size, data_layout,
                                    kernel_layout, out_layout, out_dtype)
@@ -2251,7 +2267,8 @@ def bitserial_conv2d(data,
     result : tvm.relay.Expr
         The computed result.
     """
-
+    # convert 2-way padding to 4-way padding
+    padding = get_pad_tuple2d(padding)
     return _make.bitserial_conv2d(data, weight, strides, padding, channels,
                                   kernel_size, activation_bits, weight_bits,
                                   data_layout, kernel_layout, pack_dtype,

python/tvm/relay/op/nn/util.py

@@ -0,0 +1,56 @@
+# Licensed to the Apache Software Foundation (ASF) under one
+# or more contributor license agreements.  See the NOTICE file
+# distributed with this work for additional information
+# regarding copyright ownership.  The ASF licenses this file
+# to you under the Apache License, Version 2.0 (the
+# "License"); you may not use this file except in compliance
+# with the License.  You may obtain a copy of the License at
+#
+#   http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing,
+# software distributed under the License is distributed on an
+# "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
+# KIND, either express or implied.  See the License for the
+# specific language governing permissions and limitations
+# under the License.
+# pylint: disable=invalid-name, unused-variable
+"""NN operator common utilities"""
+from __future__ import absolute_import
+from .... import container
+
+def get_pad_tuple2d(padding):
+    """Common code to get the pad option
+    Parameters
+    ----------
+    padding : Union[int, Tuple[int, ...]]
+        Padding size
+    Returns
+    -------
+    pad_top : int
+        Padding size on top
+    pad_left : int
+        Padding size on left
+    pad_down : int
+        Padding size on down.
+    pad_right : int
+        Padding size on right.
+    """
+    # compute the padding size
+    if isinstance(padding, container.Array):
+        padding = list(padding)
+    if isinstance(padding, (tuple, list)):
+        if len(padding) == 2:
+            pad_h = padding[0] * 2
+            pad_w = padding[1] * 2
+        elif len(padding) == 4:
+            return padding[0], padding[1], padding[2], padding[3]
+        else:
+            raise ValueError("Size of padding can only be 2 or 4")
+    elif isinstance(padding, int):
+        pad_h = pad_w = padding * 2
+    else:
+        raise ValueError("Unknown padding option %s" % padding)
+    pad_top = (pad_h + 1) // 2
+    pad_left = (pad_w + 1) // 2
+    return pad_top, pad_left, pad_h - pad_top, pad_w - pad_left

tests/python/relay/test_pass_alter_op_layout.py

@@ -617,7 +617,7 @@ def expected():
         w = relay.var("w", shape=(32, 1, 3, 3))
         x = relay.layout_transform(x, "NCHW", "NCHW8c")
         w = relay.layout_transform(w, "OIHW", "OIHW1i8o")
-        y = relay.nn.contrib_depthwise_conv2d_nchwc(x, w, padding=(1, 1), channels=32, kernel_size=(3, 3),
+        y = relay.nn.contrib_depthwise_conv2d_nchwc(x, w, padding=(1, 1, 1, 1), channels=32, kernel_size=(3, 3),
                                                     groups=32, data_layout="NCHW8c", kernel_layout="OIHW1i8o",
                                                     out_layout="NCHW8c")
         y = relay.layout_transform(y, "NCHW8c", "NCHW")

tests/python/unittest/test_graph_tuner_core.py

@@ -50,9 +50,9 @@ def _create_data(target, dshape, dtype, layout):
                                               params=params,
                                               ops=(relay.op.nn.conv2d,))
     wkl_list = [
-        create_workload((1, 3, 8, 8), (16, 3, 3, 3), (1, 1), (1, 1), (1, 1), layout, layout, dtype, dtype),
-        create_workload((1, 16, 8, 8), (32, 16, 1, 1), (1, 1), (0, 0), (1, 1), layout, layout, dtype, dtype),
-        create_workload((1, 32, 8, 8), (32, 32, 3, 3), (1, 1), (1, 1), (1, 1), layout, layout, dtype, dtype),
+        create_workload((1, 3, 8, 8), (16, 3, 3, 3), (1, 1), (1, 1, 1, 1), (1, 1), layout, layout, dtype, dtype),
+        create_workload((1, 16, 8, 8), (32, 16, 1, 1), (1, 1), (0, 0, 0, 0), (1, 1), layout, layout, dtype, dtype),
+        create_workload((1, 32, 8, 8), (32, 32, 3, 3), (1, 1), (1, 1, 1, 1), (1, 1), layout, layout, dtype, dtype),
     ]
     costs = [0.04, 0.012, 0.03]
     config_list = []
@@ -279,9 +279,9 @@ def test_many_sub_graphs():
                                               params=params,
                                               ops=(relay.op.nn.conv2d,))
     wkl_list = [
-        create_workload((1, 3, 8, 8), (16, 3, 3, 3), (1, 1), (1, 1), (1, 1), layout, layout, dtype, dtype),
-        create_workload((1, 16, 8, 8), (32, 16, 1, 1), (1, 1), (0, 0), (1, 1), layout, layout, dtype, dtype),
-        create_workload((1, 32, 8, 8), (32, 32, 3, 3), (1, 1), (1, 1), (1, 1), layout, layout, dtype, dtype),
+        create_workload((1, 3, 8, 8), (16, 3, 3, 3), (1, 1), (1, 1, 1, 1), (1, 1), layout, layout, dtype, dtype),
+        create_workload((1, 16, 8, 8), (32, 16, 1, 1), (1, 1), (0, 0, 0, 0), (1, 1), layout, layout, dtype, dtype),
+        create_workload((1, 32, 8, 8), (32, 32, 3, 3), (1, 1), (1, 1, 1, 1), (1, 1), layout, layout, dtype, dtype),
     ]
     costs = [0.04, 0.012, 0.03, 0.02, 0.02, 0.045]
     config_list = []
@@ -392,8 +392,8 @@ def test_tuple():
                                               params=params,
                                               ops=(relay.op.nn.conv2d,))
     wkl_list = [
-        create_workload((1, 5, 32, 32), (2, 5, 3, 3), (1, 1), (1, 1), (1, 1), layout, layout, dtype, dtype),
-        create_workload((1, 5, 32, 32), (3, 5, 3, 3), (1, 1), (1, 1), (1, 1), layout, layout, dtype, dtype),
+        create_workload((1, 5, 32, 32), (2, 5, 3, 3), (1, 1), (1, 1, 1, 1), (1, 1), layout, layout, dtype, dtype),
+        create_workload((1, 5, 32, 32), (3, 5, 3, 3), (1, 1), (1, 1, 1, 1), (1, 1), layout, layout, dtype, dtype),
     ]
     costs = [0.01, 0.012, 0.03, 0.04]
     config_list = []
@@ -490,9 +490,9 @@ def test_triangle_block():
                                               params=params,
                                               ops=(relay.op.nn.conv2d,))
     wkl_list = [
-        create_workload((1, 3, 8, 8), (16, 3, 3, 3), (1, 1), (1, 1), (1, 1), layout, layout, dtype, dtype),
-        create_workload((1, 16, 8, 8), (32, 16, 1, 1), (1, 1), (0, 0), (1, 1), layout, layout, dtype, dtype),
-        create_workload((1, 3, 8, 8), (32, 3, 3, 3), (1, 1), (1, 1), (1, 1), layout, layout, dtype, dtype),
+        create_workload((1, 3, 8, 8), (16, 3, 3, 3), (1, 1), (1, 1, 1, 1), (1, 1), layout, layout, dtype, dtype),
+        create_workload((1, 16, 8, 8), (32, 16, 1, 1), (1, 1), (0, 0, 0, 0), (1, 1), layout, layout, dtype, dtype),
+        create_workload((1, 3, 8, 8), (32, 3, 3, 3), (1, 1), (1, 1, 1, 1), (1, 1), layout, layout, dtype, dtype),
     ]
     costs = [0.04, 0.012, 0.03, 0.02, 0.02, 0.045]
     config_list = []

topi/python/topi/cuda/conv2d.py

@@ -85,7 +85,8 @@ def conv2d_cuda(cfg, data, kernel, strides, padding, dilation, layout='NCHW', ou
         stride_h, stride_w = (strides, strides) if isinstance(strides, int) else strides
         dilation_h, dilation_w = (dilation, dilation) if isinstance(dilation, int) else dilation
 
-        if isinstance(padding, (list, tuple)) and len(padding) > 2:
+        if isinstance(padding, (list, tuple)) and len(padding) == 4 and \
+           (padding[0] != padding[2] or padding[1] != padding[3]):
             raise ValueError("Cudnn doesn't support asymmetric padding.")
         pt, pl, pb, pr = get_pad_tuple(padding, (KH, KW))
         OH = (H + pt + pb - KH) // stride_h + 1