for release-1.1 (#569)

* Update pull_request_template.md

* Update pull_request_template.md

* Update pull_request_template.md

* Update README.md

* fix the pad in tf  (#567)

* fix the convert.py args

* fix teh pad and add log_softmax

* Update pull_request_template.md

* Update pull_request_template.md

* Update pull_request_template.md
Co-authored-by: NJason <jiangjiajun@baidu.com>

.github/pull_request_template.md

 # Create A Good Pull Request
 
-- [] Python code style verification
-- [] All models(TensorFLow/Caffe/ONNX/PyTorch) testing passed
-- [] Details about your pull request, releated issues
+> 下面的文字请保留在PR说明的最后面，并在提完PR后，根据实际情况勾选确认以下情况  
+
+Please check the follow step before merging this pull request
+
+- [ ] Python code style verification
+- [ ] Review all the code diff by yourself
+- [ ] All models(TensorFLow/Caffe/ONNX/PyTorch) testing passed
+- [ ] Details about your pull request, releated issues
+
+If this PR add new model support, please update `model_zoo.md` and add model to out test model zoos(@wjj19950828)
+- [ ] New Model Supported
+- [ ] No New Model Supported

README.md

@@ -116,9 +116,9 @@ X2Paddle提供了工具解决如下问题，详见[tools/README.md](tools/README
 1. 新增支持转换的模型：[SwinTransformer](https://github.com/microsoft/Swin-Transformer/)、[BASNet](https://github.com/xuebinqin/BASNet)、[DBFace](https://github.com/dlunion/DBFace)、[EasyOCR](https://github.com/JaidedAI/EasyOCR)、[CifarNet](https://github.com/tensorflow/models/blob/master/research/slim/nets/cifarnet.py)等。
 2. 支持Windows上使用本工具。
 3. 新增TensorFlow op映射（4个）：SplitV、ReverseV2、BatchToSpaceND、SpaceToBatchND。
-4. 新增PyTorch op映射（11个）：aten::index、aten::roll、aten::adaptive_avg_pool1d、aten::reflection_pad2d、aten::reflection_pad1d、aten::instance_norm、aten::gru、aten::norm、aten::clamp_min、aten:prelu、aten:split_with_sizes。
+4. 新增PyTorch op映射（11个）：aten::index、aten::roll、aten::adaptive_avg_pool1d、aten::reflection_pad2d、aten::reflection_pad1d、aten::instance_norm、aten::gru、aten::norm、aten::clamp_min、aten::prelu、aten:split_with_sizes。
 5. 新增ONNX op映射（1个）：DepthToSpace。
-6. 新增Caffe op映射（1个）：op：MemoryData。
+6. 新增Caffe op映射（1个）：MemoryData。
 
 ## 贡献代码
 

x2paddle/op_mapper/dygraph/pytorch2paddle/aten.py

@@ -2962,6 +2962,51 @@ def aten_log(mapper, graph, node):
     return current_inputs, current_outputs
 
 
+def aten_log_softmax(mapper, graph, node):
+    """ 构造log_softmax的PaddleLayer。
+    TorchScript示例:
+        %4 = aten::log_softmax(%input, %2, %3)
+        参数含义:
+        %4 (Tensor): 输出的Tensor。
+        %input (Tensor): 输入的Tensor。
+        %2 (int): 指定对输入进行运算的轴。
+        %3 (int): 输入Tensor的数据类型。
+    """
+    scope_name = mapper.normalize_scope_name(node)
+    output_name = mapper._get_outputs_name(node)[0]
+    layer_outputs = [output_name]
+    layer_inputs = {}
+    layer_attrs = {}
+    inputs_name, inputs_node = mapper._get_inputs_name(node)
+    # 获取当前节点输出的list
+    current_outputs = [output_name]
+    current_inputs = []
+    # 处理输入0，即%input
+    mapper._check_input(graph, inputs_node[0], inputs_name[0], current_outputs,
+                        scope_name)
+    layer_inputs["x"] = inputs_name[0]
+    # 处理输入1，即%2，代表dtype
+    if inputs_name[1] in mapper.attrs:
+        layer_attrs["axis"] = mapper.attrs[inputs_name[1]]
+    else:
+        mapper._check_input(graph, inputs_node[1], inputs_name[1],
+                            current_outputs, scope_name)
+        layer_inputs["axis"] = inputs_name[1]
+    # 处理输入2，即%3，代表dtype
+    if mapper.attrs[inputs_name[2]] is not None:
+        layer_attrs["dtype"] = dtype_dict[mapper.attrs[inputs_name[2]]]
+    # 获取当前节点输入的list
+    current_inputs = list(layer_inputs.values())
+
+    graph.add_layer(
+        "paddle.nn.functional.log_softmax",
+        inputs=layer_inputs,
+        outputs=layer_outputs,
+        scope_name=scope_name,
+        **layer_attrs)
+    return current_inputs, current_outputs
+
+
 def aten_lstm(mapper, graph, node):
     """ 构造长短期记忆网络（LSTM）的PaddleLayer。
     TorchScript示例:

x2paddle/op_mapper/dygraph/tf2paddle/tf_op_mapper.py

@@ -649,18 +649,30 @@ class TFOpMapper(OpMapper):
         paddings = self.graph.get_input_node(node, 1)
         assert paddings.layer_type == "Const", "Padding should be Const"
         paddings = paddings.value.flatten().tolist()
+            
         constant_values = 0
         if len(node.layer.input) > 2:
             constant_values = self.graph.get_input_node(node, 2)
             assert constant_values.layer_type == "Const", "Padding should be Const"
             constant_values = constant_values.value
-
-        self.paddle_graph.add_layer(
-            kernel="paddle.nn.functional.pad",
-            inputs={"x": input.name},
-            outputs=[node.name],
-            pad=paddings,
-            value=constant_values)
+            
+        if len(paddings) == 8 and sum(paddings[:2]) == 0 \
+            and sum(paddings[-2:]) == 0:
+            paddings = paddings[2: -2]
+            self.paddle_graph.add_layer(
+                kernel="paddle.nn.functional.pad",
+                inputs={"x": input.name},
+                outputs=[node.name],
+                pad=paddings,
+                value=constant_values,
+                data_format=string('NHWC'))
+        else:
+            self.paddle_graph.add_layer(
+                kernel="paddle.nn.functional.pad",
+                inputs={"x": input.name},
+                outputs=[node.name],
+                pad=paddings,
+                value=constant_values)
 
     def MirrorPad(self, node):
         self.Pad(node)