add conv2d op

examples/oneflow2onnx/nodes/test_bias_add.py

-"""
-Copyright 2020 The OneFlow Authors. All rights reserved.
-
-Licensed 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.
-"""
-import oneflow as flow
-import oneflow.typing as tp
-from oneflow_onnx.oneflow2onnx.util import convert_to_onnx_and_check
-
-
-def test_bias_add_nchw():
-    @flow.global_function()
-    def bias_add_nchw(x: tp.Numpy.Placeholder((3, 4, 2, 5))):
-        y = flow.get_variable(
-            name="y",
-            shape=(4,),
-            dtype=flow.float,
-            initializer=flow.random_uniform_initializer(),
-        )
-        return flow.nn.bias_add(x, y, "NCHW")
-
-    convert_to_onnx_and_check(bias_add_nchw)
-
-
-def test_bias_add_nhwc():
-    @flow.global_function()
-    def bias_add_nhwc(x: tp.Numpy.Placeholder((3, 4, 2, 5))):
-        y = flow.get_variable(
-            name="y",
-            shape=(5,),
-            dtype=flow.float,
-            initializer=flow.random_uniform_initializer(),
-        )
-        return flow.nn.bias_add(x, y, "NHWC")
-
-    convert_to_onnx_and_check(bias_add_nhwc)

examples/oneflow2onnx/nodes/test_concat.py

@@ -13,42 +13,35 @@ 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.
 """
+import tempfile
 import oneflow as flow
-import oneflow.typing as tp
 from oneflow_onnx.oneflow2onnx.util import convert_to_onnx_and_check
 
+class ConCat(flow.nn.Module):
+    def __init__(self) -> None:
+        super(ConCat, self).__init__()
+    
+    def forward(self, x: flow.Tensor) -> flow.Tensor:
+        return flow.cat([x, x, x], dim=1)
 
-def test_concat_axis0():
-    @flow.global_function()
-    def concat():
-        variables = []
-        for i in range(4):
-            variables.append(
-                flow.get_variable(
-                    name=str(i),
-                    shape=(2, 3),
-                    dtype=flow.float,
-                    initializer=flow.random_uniform_initializer(),
-                )
-            )
-        return flow.concat(variables, axis=0)
-
-    convert_to_onnx_and_check(concat)
-
-
-def test_concat_axis1():
-    @flow.global_function()
-    def concat():
-        variables = []
-        for i in range(4):
-            variables.append(
-                flow.get_variable(
-                    name=str(i),
-                    shape=(2, 3),
-                    dtype=flow.float,
-                    initializer=flow.random_uniform_initializer(),
-                )
-            )
-        return flow.concat(variables, axis=1)
-
-    convert_to_onnx_and_check(concat)
+concat = ConCat()
+class ConCatOpGraph(flow.nn.Graph):
+    def __init__(self):
+        super().__init__()
+        self.m = concat
+
+    def build(self, x):
+        out = self.m(x)
+        return out
+
+
+def test_concat():
+    
+    concat_graph = ConCatOpGraph()
+    concat_graph._compile(flow.randn(1, 3, 224, 224))
+
+    with tempfile.TemporaryDirectory() as tmpdirname:
+        flow.save(concat.state_dict(), tmpdirname)
+        convert_to_onnx_and_check(concat_graph, flow_weight_dir=tmpdirname, onnx_model_path="/tmp")
+
+test_concat()

examples/oneflow2onnx/nodes/test_constant.py

@@ -13,21 +13,35 @@ 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.
 """
+import tempfile
 import oneflow as flow
-import oneflow.typing as tp
 from oneflow_onnx.oneflow2onnx.util import convert_to_onnx_and_check
 
+class Constant(flow.nn.Module):
+    def __init__(self) -> None:
+        super(Constant, self).__init__()
+    
+    def forward(self, x: flow.Tensor) -> flow.Tensor:
+        return flow.ones((2, 3)) + flow.zeros((2, 3))
 
-def test_constant_float():
-    @flow.global_function()
-    def constant(x: tp.Numpy.Placeholder((3, 5))):
-        return flow.constant(value=1.5, shape=(1, 3, 3), dtype=flow.float)
+constant = Constant()
+class constantOpGraph(flow.nn.Graph):
+    def __init__(self):
+        super().__init__()
+        self.m = constant
 
-    convert_to_onnx_and_check(constant)
+    def build(self, x):
+        out = self.m(x)
+        return out
 
-def test_constant_int():
-    @flow.global_function()
-    def constant(x: tp.Numpy.Placeholder((3, 5))):
-        return flow.constant(value=1, shape=(1, 3, 3), dtype=flow.int)
 
-    convert_to_onnx_and_check(constant)
\ No newline at end of file
+def test_constant():
+    
+    constant_graph = constantOpGraph()
+    constant_graph._compile(flow.randn(1, 3, 224, 224))
+
+    with tempfile.TemporaryDirectory() as tmpdirname:
+        flow.save(constant.state_dict(), tmpdirname)
+        convert_to_onnx_and_check(constant_graph, flow_weight_dir=tmpdirname, onnx_model_path="/tmp")
+
+test_constant()

examples/oneflow2onnx/nodes/test_conv.py

@@ -13,106 +13,36 @@ 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.
 """
+import tempfile
 import oneflow as flow
-import oneflow.typing as tp
 from oneflow_onnx.oneflow2onnx.util import convert_to_onnx_and_check
 
-initializer = flow.random_uniform_initializer()
-initer_args = {"kernel_initializer": initializer, "bias_initializer": initializer}
+class Conv2d(flow.nn.Module):
+    def __init__(self) -> None:
+        super(Conv2d, self).__init__()
+        self.conv = flow.nn.Conv2d(3, 16, 3)
 
+    def forward(self, x: flow.Tensor) -> flow.Tensor:
+        return self.conv(x)
 
-def test_conv2d_k2d1_valid():
-    @flow.global_function()
-    def conv2d_k3s1_valid(x: tp.Numpy.Placeholder((2, 4, 3, 5))):
-        return flow.layers.conv2d(
-            x, 6, kernel_size=3, strides=1, padding="VALID", **initer_args
-        )
+conv = Conv2d()
+class convOpGraph(flow.nn.Graph):
+    def __init__(self):
+        super().__init__()
+        self.m = conv
 
-    convert_to_onnx_and_check(conv2d_k3s1_valid)
+    def build(self, x):
+        out = self.m(x)
+        return out
 
 
-def test_conv2d_s2_valid():
-    @flow.global_function()
-    def conv2d_s2_valid(x: tp.Numpy.Placeholder((2, 4, 3, 5))):
-        return flow.layers.conv2d(
-            x, 6, kernel_size=1, strides=2, padding="VALID", **initer_args
-        )
+def test_conv():
+    
+    conv_graph = convOpGraph()
+    conv_graph._compile(flow.randn(1, 3, 224, 224))
 
-    convert_to_onnx_and_check(conv2d_s2_valid)
+    with tempfile.TemporaryDirectory() as tmpdirname:
+        flow.save(conv.state_dict(), tmpdirname)
+        convert_to_onnx_and_check(conv_graph, flow_weight_dir=tmpdirname, onnx_model_path="/tmp")
 
-
-def test_conv2d_s2_same():
-    @flow.global_function()
-    def conv2d_s2_same(x: tp.Numpy.Placeholder((2, 4, 3, 5))):
-        return flow.layers.conv2d(
-            x, 6, kernel_size=3, strides=2, padding="SAME", **initer_args
-        )
-
-    convert_to_onnx_and_check(conv2d_s2_same)
-
-
-def test_conv2d_k3s1_nhwc_valid():
-    @flow.global_function()
-    def conv2d_k3s1_nhwc_valid(x: tp.Numpy.Placeholder((2, 3, 5, 4))):
-        return flow.layers.conv2d(
-            x,
-            6,
-            kernel_size=3,
-            strides=1,
-            padding="VALID",
-            data_format="NHWC",
-            **initer_args
-        )
-
-    convert_to_onnx_and_check(conv2d_k3s1_nhwc_valid)
-
-
-def test_conv2d_k3s1_nhwc_same_d2():
-    @flow.global_function()
-    def conv2d(x: tp.Numpy.Placeholder((2, 7, 5, 4))):
-        return flow.layers.conv2d(
-            x,
-            6,
-            kernel_size=3,
-            strides=1,
-            dilation_rate=2,
-            padding="SAME",
-            data_format="NHWC",
-            **initer_args
-        )
-
-    convert_to_onnx_and_check(conv2d)
-
-
-def test_conv2d_k3s1_nchw_same_g2():
-    @flow.global_function()
-    def conv2d(x: tp.Numpy.Placeholder((2, 4, 5, 3))):
-        return flow.layers.conv2d(
-            x,
-            6,
-            kernel_size=3,
-            strides=1,
-            groups=2,
-            padding="SAME",
-            data_format="NCHW",
-            **initer_args
-        )
-
-    convert_to_onnx_and_check(conv2d)
-
-
-def test_conv2d_k3s1_nchw_same_depthwise():
-    @flow.global_function()
-    def conv2d(x: tp.Numpy.Placeholder((2, 4, 5, 3))):
-        return flow.layers.conv2d(
-            x,
-            4,
-            kernel_size=3,
-            strides=1,
-            groups=4,
-            padding="SAME",
-            data_format="NCHW",
-            **initer_args
-        )
-
-    convert_to_onnx_and_check(conv2d)
+test_conv()