CINN sub graph test - 6

test/ir/pir/cinn/sub_graphs/test_sub_graph_3.py

@@ -48,8 +48,8 @@ def forward(
         var_9 = var_8.reshape((0, 16, 49))
         var_10 = paddle.tensor.linalg.transpose(var_1, perm=[0, 1, 3, 2])
         var_11 = paddle.tensor.linalg.matmul(var_5, var_10)
-        var_12 = var_11.__mul__(0.25)
-        var_13 = var_12.__add__(var_9)
+        var_12 = var_11 * 0.25
+        var_13 = var_12 + var_9
         var_14 = paddle.nn.functional.activation.softmax(var_13)
         var_15 = paddle.tensor.linalg.matmul(var_14, var_2)
         var_16 = paddle.tensor.linalg.transpose(var_15, perm=[0, 2, 1, 3])

test/ir/pir/cinn/sub_graphs/test_sub_graph_31.py

@@ -0,0 +1,80 @@
+# Copyright (c) 2024 PaddlePaddle 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.
+
+# repo: PaddleClas
+# model: ppcls^configs^ImageNet^DPN^DPN98
+# api:paddle.tensor.manipulation.split||api:paddle.tensor.math.add||api:paddle.tensor.manipulation.concat
+import unittest
+
+import numpy as np
+
+import paddle
+
+
+class LayerCase(paddle.nn.Layer):
+    def __init__(self):
+        super().__init__()
+
+    def forward(
+        self,
+        var_0,  # (shape: [22, 1056, 14, 14], dtype: paddle.float32, stop_gradient: False)
+        var_1,  # (shape: [22, 1024, 14, 14], dtype: paddle.float32, stop_gradient: False)
+        var_2,  # (shape: [22, 288, 14, 14], dtype: paddle.float32, stop_gradient: False)
+    ):
+        var_3, var_4 = paddle.tensor.manipulation.split(
+            var_0, num_or_sections=[1024, 32], axis=1
+        )
+        var_5 = paddle.tensor.math.add(x=var_1, y=var_3)
+        var_6 = paddle.tensor.manipulation.concat([var_2, var_4], axis=1)
+        return var_5, var_6
+
+
+class TestLayer(unittest.TestCase):
+    def setUp(self):
+        self.inputs = (
+            paddle.rand(shape=[22, 1056, 14, 14], dtype=paddle.float32),
+            paddle.rand(shape=[22, 1024, 14, 14], dtype=paddle.float32),
+            paddle.rand(shape=[22, 288, 14, 14], dtype=paddle.float32),
+        )
+        self.net = LayerCase()
+
+    def train(self, net, to_static, with_prim=False, with_cinn=False):
+        if to_static:
+            paddle.set_flags({'FLAGS_prim_all': with_prim})
+            if with_cinn:
+                build_strategy = paddle.static.BuildStrategy()
+                build_strategy.build_cinn_pass = True
+                net = paddle.jit.to_static(
+                    net, build_strategy=build_strategy, full_graph=True
+                )
+            else:
+                net = paddle.jit.to_static(net, full_graph=True)
+        paddle.seed(123)
+        outs = net(*self.inputs)
+        return outs
+
+    # NOTE prim + cinn lead to error
+    def test_ast_prim_cinn(self):
+        st_out = self.train(self.net, to_static=True)
+        cinn_out = self.train(
+            self.net, to_static=True, with_prim=True, with_cinn=False
+        )
+        for st, cinn in zip(
+            paddle.utils.flatten(st_out), paddle.utils.flatten(cinn_out)
+        ):
+            np.testing.assert_allclose(st.numpy(), cinn.numpy(), atol=1e-8)
+
+
+if __name__ == '__main__':
+    unittest.main()

test/ir/pir/cinn/sub_graphs/test_sub_graph_32.py

@@ -0,0 +1,79 @@
+# Copyright (c) 2024 PaddlePaddle 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.
+
+# repo: PaddleClas
+# model: ppcls^configs^ImageNet^ResNeSt^ResNeSt50
+# api:paddle.tensor.manipulation.reshape||api:paddle.tensor.linalg.transpose||api:paddle.nn.functional.activation.softmax||api:paddle.tensor.manipulation.reshape
+import unittest
+
+import numpy as np
+
+import paddle
+
+
+class LayerCase(paddle.nn.Layer):
+    def __init__(self):
+        super().__init__()
+
+    def forward(
+        self,
+        var_0,  # (shape: [22, 1024, 1, 1], dtype: paddle.float32, stop_gradient: True)
+    ):
+        var_1 = paddle.tensor.manipulation.reshape(
+            x=var_0, shape=[22, 1, 2, 512]
+        )
+        var_2 = paddle.tensor.linalg.transpose(x=var_1, perm=[0, 2, 1, 3])
+        var_3 = paddle.nn.functional.activation.softmax(var_2, axis=1)
+        var_4 = paddle.tensor.manipulation.reshape(
+            x=var_3, shape=[22, 1024, 1, 1]
+        )
+        return var_4
+
+
+class TestLayer(unittest.TestCase):
+    def setUp(self):
+        self.inputs = (
+            paddle.rand(shape=[22, 1024, 1, 1], dtype=paddle.float32),
+        )
+        self.net = LayerCase()
+
+    def train(self, net, to_static, with_prim=False, with_cinn=False):
+        if to_static:
+            paddle.set_flags({'FLAGS_prim_all': with_prim})
+            if with_cinn:
+                build_strategy = paddle.static.BuildStrategy()
+                build_strategy.build_cinn_pass = True
+                net = paddle.jit.to_static(
+                    net, build_strategy=build_strategy, full_graph=True
+                )
+            else:
+                net = paddle.jit.to_static(net, full_graph=True)
+        paddle.seed(123)
+        outs = net(*self.inputs)
+        return outs
+
+    # NOTE prim + cinn lead to error
+    def test_ast_prim_cinn(self):
+        st_out = self.train(self.net, to_static=True)
+        cinn_out = self.train(
+            self.net, to_static=True, with_prim=True, with_cinn=False
+        )
+        for st, cinn in zip(
+            paddle.utils.flatten(st_out), paddle.utils.flatten(cinn_out)
+        ):
+            np.testing.assert_allclose(st.numpy(), cinn.numpy(), atol=1e-8)
+
+
+if __name__ == '__main__':
+    unittest.main()

test/ir/pir/cinn/sub_graphs/test_sub_graph_33.py

@@ -0,0 +1,102 @@
+# Copyright (c) 2024 PaddlePaddle 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.
+
+# repo: PaddleClas
+# model: ppcls^configs^ImageNet^RedNet^RedNet38
+# api:paddle.nn.functional.conv._conv_nd||method:reshape||method:unsqueeze||api:paddle.nn.functional.common.unfold||method:reshape||method:__mul__||method:sum||method:reshape
+import unittest
+
+import numpy as np
+
+import paddle
+
+
+class LayerCase(paddle.nn.Layer):
+    def __init__(self):
+        super().__init__()
+        self.parameter_0 = self.create_parameter(
+            shape=[784],
+            dtype=paddle.float32,
+        )
+        self.parameter_1 = self.create_parameter(
+            shape=[784, 64, 1, 1],
+            dtype=paddle.float32,
+        )
+
+    def forward(
+        self,
+        var_0,  # (shape: [10, 64, 14, 14], dtype: paddle.float32, stop_gradient: False)
+        var_1,  # (shape: [10, 256, 14, 14], dtype: paddle.float32, stop_gradient: False)
+    ):
+        var_2 = paddle.nn.functional.conv._conv_nd(
+            var_0,
+            self.parameter_1,
+            bias=self.parameter_0,
+            stride=[1, 1],
+            padding=[0, 0],
+            padding_algorithm='EXPLICIT',
+            dilation=[1, 1],
+            groups=1,
+            data_format='NCHW',
+            channel_dim=1,
+            op_type='conv2d',
+            use_cudnn=True,
+        )
+        var_3 = var_2.reshape((10, 16, 49, 14, 14))
+        var_4 = var_3.unsqueeze(2)
+        var_5 = paddle.nn.functional.common.unfold(var_1, 7, 1, 3, 1)
+        var_6 = var_5.reshape((10, 16, 16, 49, 14, 14))
+        var_7 = var_4 * var_6
+        var_8 = var_7.sum(axis=3)
+        var_9 = var_8.reshape((10, 256, 14, 14))
+        return var_9
+
+
+class TestLayer(unittest.TestCase):
+    def setUp(self):
+        self.inputs = (
+            paddle.rand(shape=[10, 64, 14, 14], dtype=paddle.float32),
+            paddle.rand(shape=[10, 256, 14, 14], dtype=paddle.float32),
+        )
+        self.net = LayerCase()
+
+    def train(self, net, to_static, with_prim=False, with_cinn=False):
+        if to_static:
+            paddle.set_flags({'FLAGS_prim_all': with_prim})
+            if with_cinn:
+                build_strategy = paddle.static.BuildStrategy()
+                build_strategy.build_cinn_pass = True
+                net = paddle.jit.to_static(
+                    net, build_strategy=build_strategy, full_graph=True
+                )
+            else:
+                net = paddle.jit.to_static(net, full_graph=True)
+        paddle.seed(123)
+        outs = net(*self.inputs)
+        return outs
+
+    # NOTE prim + cinn lead to error
+    def test_ast_prim_cinn(self):
+        st_out = self.train(self.net, to_static=True)
+        cinn_out = self.train(
+            self.net, to_static=True, with_prim=True, with_cinn=False
+        )
+        for st, cinn in zip(
+            paddle.utils.flatten(st_out), paddle.utils.flatten(cinn_out)
+        ):
+            np.testing.assert_allclose(st.numpy(), cinn.numpy(), atol=1e-8)
+
+
+if __name__ == '__main__':
+    unittest.main()

test/ir/pir/cinn/sub_graphs/test_sub_graph_34.py

@@ -0,0 +1,76 @@
+# Copyright (c) 2024 PaddlePaddle 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.
+
+# repo: PaddleClas
+# model: ppcls^configs^ImageNet^HRNet^HRNet_W32_C
+# api:paddle.nn.functional.common.upsample||api:paddle.tensor.math.add
+import unittest
+
+import numpy as np
+
+import paddle
+
+
+class LayerCase(paddle.nn.Layer):
+    def __init__(self):
+        super().__init__()
+
+    def forward(
+        self,
+        var_0,  # (shape: [10, 32, 14, 14], dtype: paddle.float32, stop_gradient: False)
+        var_1,  # (shape: [10, 32, 56, 56], dtype: paddle.float32, stop_gradient: False)
+    ):
+        var_2 = paddle.nn.functional.common.upsample(
+            var_0, scale_factor=4, mode='nearest'
+        )
+        var_3 = paddle.tensor.math.add(x=var_1, y=var_2)
+        return var_2, var_3
+
+
+class TestLayer(unittest.TestCase):
+    def setUp(self):
+        self.inputs = (
+            paddle.rand(shape=[10, 32, 14, 14], dtype=paddle.float32),
+            paddle.rand(shape=[10, 32, 56, 56], dtype=paddle.float32),
+        )
+        self.net = LayerCase()
+
+    def train(self, net, to_static, with_prim=False, with_cinn=False):
+        if to_static:
+            paddle.set_flags({'FLAGS_prim_all': with_prim})
+            if with_cinn:
+                build_strategy = paddle.static.BuildStrategy()
+                build_strategy.build_cinn_pass = True
+                net = paddle.jit.to_static(
+                    net, build_strategy=build_strategy, full_graph=True
+                )
+            else:
+                net = paddle.jit.to_static(net, full_graph=True)
+        paddle.seed(123)
+        outs = net(*self.inputs)
+        return outs
+
+    def test_ast_prim_cinn(self):
+        st_out = self.train(self.net, to_static=True)
+        cinn_out = self.train(
+            self.net, to_static=True, with_prim=True, with_cinn=True
+        )
+        for st, cinn in zip(
+            paddle.utils.flatten(st_out), paddle.utils.flatten(cinn_out)
+        ):
+            np.testing.assert_allclose(st.numpy(), cinn.numpy(), atol=1e-8)
+
+
+if __name__ == '__main__':
+    unittest.main()