add op:swish,floor,uniform_random_batch_size_like

x2paddle/op_mapper/paddle2onnx/opset9/opset.py

@@ -405,6 +405,23 @@ class OpSet9(object):
             'Sum', inputs=op.input('X'), outputs=op.output('Out'))
         return node
 
+    def floor(self, op, block):
+        node = helper.make_node(
+            'Floor', inputs=op.input('X'), outputs=op.output('Out'))
+        return node
+
+    def uniform_random_batch_size_like(self, op, block):
+        node = helper.make_node(
+            'RandomUniformLike',
+            inputs=op.input('Input'),
+            outputs=op.output('Out'),
+            #shape=op.attr('shape'),
+            high=op.attr('max'),
+            dtype=self.paddle_onnx_dtype_map[op.attr('dtype')],
+            low=op.attr('min'),
+            seed=float(op.attr('seed')), )
+        return node
+
     def depthwise_conv2d(self, op, block):
         return self.conv2d(op, block)
 
@@ -784,6 +801,38 @@ class OpSet9(object):
             beta=offset)
         return node
 
+    def swish(self, op, block):
+        beta = op.attr('beta')
+        beta_name = self.get_name(op.type, 'beta')
+        beta_node = onnx.helper.make_node(
+            'Constant',
+            name=beta_name,
+            inputs=[],
+            outputs=[beta_name],
+            value=onnx.helper.make_tensor(
+                name=beta_name,
+                data_type=onnx.TensorProto.FLOAT,
+                dims=(),
+                vals=[beta]))
+
+        beta_x_name = self.get_name(op.type, 'beta_x')
+        beta_x_node = onnx.helper.make_node(
+            'Mul',
+            name=beta_x_name,
+            inputs=[op.input('X')[0], beta_name],
+            outputs=[beta_x_name])
+        sigmoid_name = self.get_name(op.type, 'sigmoid')
+        sigmoid_node = onnx.helper.make_node(
+            'Sigmoid',
+            name=sigmoid_name,
+            inputs=[beta_x_name],
+            outputs=[sigmoid_name])
+        swish_node = onnx.helper.make_node(
+            'Mul',
+            inputs=[op.input('X')[0], sigmoid_name],
+            outputs=op.output('Out'))
+        return [beta_node, beta_x_node, sigmoid_node, swish_node]
+
     def hard_swish(self, op, block):
         scale_name = self.get_name(op.type, 'scale')
         offset_name = self.get_name(op.type, 'offset')