increase parser st

tests/depends/ops_stub/ops_stub.h

@@ -108,6 +108,183 @@ REG_OP(PartitionedCall)
     .ATTR(config_proto, String, "")
     .ATTR(executor_type, String, "")
     .OP_END_FACTORY_REG(PartitionedCall)
+REG_OP(TensorArray)
+    .INPUT(size, TensorType({DT_INT32}))
+    .OUTPUT(handle, TensorType({DT_RESOURCE}))
+    .OUTPUT(flow, TensorType({DT_FLOAT}))
+    .REQUIRED_ATTR(dtype, Type)
+    .ATTR(element_shape, ListInt, ge::UNKNOWN_RANK)
+    .ATTR(dynamic_size, Bool, false)
+    .ATTR(clear_after_read, Bool, true)
+    .ATTR(identical_element_shapes, Bool, false)
+    .ATTR(tensor_array_name, String, "")
+    .OP_END_FACTORY_REG(TensorArray)
+
+REG_OP(TensorArrayWrite)
+    .INPUT(handle, TensorType({DT_RESOURCE}))
+    .INPUT(index, TensorType({DT_INT32}))
+    .INPUT(value, TensorType({DT_FLOAT, DT_FLOAT16, DT_DOUBLE, DT_INT8,
+        DT_INT16, DT_UINT16, DT_UINT8, DT_INT32, DT_INT64, DT_BOOL,
+        DT_STRING, DT_COMPLEX64, DT_COMPLEX128}))
+    .INPUT(flow_in, TensorType({DT_FLOAT}))
+    .OUTPUT(flow_out, TensorType({DT_FLOAT}))
+    .OP_END_FACTORY_REG(TensorArrayWrite)
+
+REG_OP(AvgPool3DGrad)
+    .INPUT(orig_input_shape, TensorType({DT_INT32}))
+    .INPUT(grads, TensorType({DT_FLOAT16, DT_FLOAT32, DT_DOUBLE}))
+    .OUTPUT(output, TensorType({DT_FLOAT16, DT_FLOAT32, DT_DOUBLE}))
+    .REQUIRED_ATTR(ksize, ListInt)
+    .REQUIRED_ATTR(strides, ListInt)
+    .REQUIRED_ATTR(pads, ListInt)
+    .ATTR(ceil_mode, Bool, false)
+    .ATTR(count_include_pad, Bool, true)
+    .ATTR(divisor_override, Int, 0)
+    .ATTR(data_format, String, "NDHWC")
+    .OP_END_FACTORY_REG(AvgPool3DGrad)
+
+REG_OP(Merge)
+    .DYNAMIC_INPUT(x, TensorType({DT_FLOAT16, DT_FLOAT, DT_DOUBLE,
+        DT_INT8, DT_INT16, DT_INT32, DT_INT64, DT_UINT8, DT_UINT16, DT_UINT32,
+        DT_UINT64, DT_BOOL}))
+    .OUTPUT(y, TensorType({DT_FLOAT16, DT_FLOAT, DT_DOUBLE,
+        DT_INT8, DT_INT16, DT_INT32, DT_INT64, DT_UINT8, DT_UINT16, DT_UINT32,
+        DT_UINT64, DT_BOOL}))
+    .OUTPUT(value_index, TensorType({DT_INT32}))
+    .OP_END_FACTORY_REG(Merge)
+
+REG_OP(NoOp)
+    .OP_END_FACTORY_REG(NoOp)
+
+REG_OP(VarIsInitializedOp)
+    .INPUT(x, TensorType({DT_FLOAT, DT_FLOAT16, DT_INT8, DT_INT16, DT_UINT16, DT_UINT8,
+                          DT_INT32, DT_INT64, DT_UINT32, DT_UINT64, DT_BOOL, DT_DOUBLE}))
+    .OUTPUT(y, TensorType({DT_BOOL}))
+    .OP_END_FACTORY_REG(VarIsInitializedOp)
+
+REG_OP(AssignVariableOp)
+    .INPUT(resource, TensorType({DT_RESOURCE}))
+    .INPUT(value, TensorType({DT_FLOAT, DT_FLOAT16, DT_INT8, DT_INT16, \
+        DT_UINT16, DT_UINT8, DT_INT32, DT_INT64, DT_BOOL, DT_DOUBLE}))
+    .REQUIRED_ATTR(dtype, Type)
+    .OP_END_FACTORY_REG(AssignVariableOp)
+
+REG_OP(ReadVariableOp)
+    .INPUT(x, TensorType({DT_FLOAT, DT_FLOAT16, DT_INT8, DT_INT16, DT_UINT16, DT_UINT8,
+                          DT_INT32, DT_INT64, DT_UINT32, DT_UINT64, DT_BOOL, DT_DOUBLE}))
+    .OUTPUT(y, TensorType({DT_FLOAT, DT_FLOAT16, DT_INT8, DT_INT16, DT_UINT16, DT_UINT8,
+                           DT_INT32, DT_INT64, DT_UINT32, DT_UINT64, DT_BOOL, DT_DOUBLE}))
+    .ATTR(dtype, Int, DT_INT32)
+    .OP_END_FACTORY_REG(ReadVariableOp)
+
+REG_OP(Reshape)
+    .INPUT(x, TensorType({DT_FLOAT, DT_FLOAT16, DT_INT8, DT_INT16, DT_UINT16, DT_UINT8, DT_INT32,
+        DT_INT64, DT_UINT32, DT_UINT64, DT_BOOL, DT_DOUBLE}))
+    .INPUT(shape, TensorType({DT_INT32, DT_INT64}))
+    .OUTPUT(y, TensorType({DT_FLOAT, DT_FLOAT16, DT_INT8, DT_INT16, DT_UINT16, DT_UINT8, DT_INT32,
+        DT_INT64, DT_UINT32, DT_UINT64, DT_BOOL, DT_DOUBLE}))
+    .ATTR(axis, Int, 0)
+    .ATTR(num_axes, Int, -1)
+    .OP_END_FACTORY_REG(Reshape)
+
+REG_OP(VarHandleOp)
+    .ATTR(container, String, "")
+    .ATTR(shared_name, String, "")
+    .REQUIRED_ATTR(dtype, Type)
+    .ATTR(shape, ListInt, ge::UNKNOWN_SHAPE)
+    .OUTPUT(y, TensorType({DT_RESOURCE}))
+    .OP_END_FACTORY_REG(VarHandleOp)
+
+REG_OP(Squeeze)
+    .INPUT(x, TensorType::ALL())
+    .OUTPUT(y, TensorType::ALL())
+    .ATTR(axis, ListInt, {})
+    .OP_END_FACTORY_REG(Squeeze)
+
+REG_OP(Fill)
+    .INPUT(dims, TensorType::IndexNumberType())
+    .INPUT(value, TensorType({DT_FLOAT, DT_DOUBLE, DT_INT32, DT_UINT8, DT_INT16,
+                              DT_INT8, DT_COMPLEX64, DT_INT64, DT_BOOL, DT_QINT8,
+                              DT_QUINT8, DT_QINT32, DT_QINT16, DT_QUINT16, DT_UINT16,
+                              DT_COMPLEX128, DT_FLOAT16, DT_UINT32, DT_UINT64}))
+    .OUTPUT(y, TensorType({DT_FLOAT, DT_DOUBLE, DT_INT32, DT_UINT8, DT_INT16,
+                              DT_INT8, DT_COMPLEX64, DT_INT64, DT_BOOL, DT_QINT8,
+                              DT_QUINT8, DT_QINT32, DT_QINT16, DT_QUINT16, DT_UINT16,
+                              DT_COMPLEX128, DT_FLOAT16, DT_UINT32, DT_UINT64}))
+    .OP_END_FACTORY_REG(Fill)
+
+REG_OP(ShapeN)
+    .DYNAMIC_INPUT(x, TensorType({DT_FLOAT, DT_FLOAT16, DT_INT8, DT_INT16, DT_UINT16, DT_UINT8,
+        DT_INT32, DT_INT64, DT_UINT32, DT_UINT64, DT_BOOL, DT_DOUBLE}))
+    .DYNAMIC_OUTPUT(y, TensorType({DT_INT32, DT_INT64}))
+    .ATTR(dtype, Int, DT_INT32)
+    .OP_END_FACTORY_REG(ShapeN)
+
+REG_OP(Switch)
+    .INPUT(data, TensorType({DT_FLOAT16, DT_FLOAT, DT_DOUBLE,
+        DT_INT8, DT_INT16, DT_INT32, DT_INT64, DT_UINT8, DT_UINT16, DT_UINT32,
+        DT_UINT64, DT_BOOL}))
+    .INPUT(pred, TensorType({DT_BOOL}))
+    .OUTPUT(output_false, TensorType({DT_FLOAT16, DT_FLOAT, DT_DOUBLE,
+        DT_INT8, DT_INT16, DT_INT32, DT_INT64, DT_UINT8, DT_UINT16, DT_UINT32,
+        DT_UINT64, DT_BOOL}))
+    .OUTPUT(output_true, TensorType({DT_FLOAT16, DT_FLOAT, DT_DOUBLE,
+        DT_INT8, DT_INT16, DT_INT32, DT_INT64, DT_UINT8, DT_UINT16, DT_UINT32,
+        DT_UINT64, DT_BOOL}))
+    .OP_END_FACTORY_REG(Switch)
+
+REG_OP(RefSwitch)
+    .INPUT(data, TensorType({DT_FLOAT16, DT_FLOAT, DT_DOUBLE,
+        DT_INT8, DT_INT16, DT_INT32, DT_INT64, DT_UINT8, DT_UINT16, DT_UINT32,
+        DT_UINT64, DT_BOOL}))
+    .INPUT(pred, TensorType({DT_BOOL}))
+    .OUTPUT(output_false, TensorType({DT_FLOAT16, DT_FLOAT, DT_DOUBLE,
+        DT_INT8, DT_INT16, DT_INT32, DT_INT64, DT_UINT8, DT_UINT16, DT_UINT32,
+        DT_UINT64, DT_BOOL}))
+    .OUTPUT(output_true, TensorType({DT_FLOAT16, DT_FLOAT, DT_DOUBLE,
+        DT_INT8, DT_INT16, DT_INT32, DT_INT64, DT_UINT8, DT_UINT16, DT_UINT32,
+        DT_UINT64, DT_BOOL}))
+    .OP_END_FACTORY_REG(RefSwitch)
+
+REG_OP(Enter)
+    .INPUT(x, TensorType({DT_FLOAT16, DT_FLOAT, DT_DOUBLE,
+        DT_INT8, DT_INT16, DT_INT32, DT_INT64, DT_UINT8, DT_UINT16, DT_UINT32,
+        DT_UINT64, DT_BOOL}))
+    .OUTPUT(y, TensorType({DT_FLOAT16, DT_FLOAT, DT_DOUBLE,
+        DT_INT8, DT_INT16, DT_INT32, DT_INT64, DT_UINT8, DT_UINT16, DT_UINT32,
+        DT_UINT64, DT_BOOL}))
+    .REQUIRED_ATTR(frame_name, String)
+    .REQUIRED_ATTR(is_constant, Bool)
+    .OP_END_FACTORY_REG(Enter)
+
+REG_OP(VariableV2)
+    .INPUT(x, TensorType({DT_FLOAT, DT_FLOAT16, DT_INT8, DT_INT16, DT_UINT16, \
+        DT_UINT8, DT_INT32, DT_INT64, DT_UINT32, DT_UINT64, DT_BOOL, DT_DOUBLE}))
+    .OUTPUT(y, TensorType({DT_FLOAT, DT_FLOAT16, DT_INT8, DT_INT16, DT_UINT16, \
+        DT_UINT8, DT_INT32, DT_INT64, DT_UINT32, DT_UINT64, DT_BOOL, DT_DOUBLE}))
+    .ATTR(index, Int, 0)
+    .ATTR(value, Tensor, Tensor())
+    .ATTR(container, String, "")
+    .ATTR(shared_name, String, "")
+    .OP_END_FACTORY_REG(VariableV2)
+
+REG_OP(Constant)
+    .OUTPUT(y, TensorType({DT_FLOAT, DT_FLOAT16, DT_INT8, DT_INT16, DT_UINT16,
+        DT_UINT8, DT_INT32, DT_INT64, DT_UINT32, DT_UINT64, DT_BOOL, DT_DOUBLE}))
+    .ATTR(value, Tensor, Tensor())
+    .OP_END_FACTORY_REG(Constant)
+
+REG_OP(Mul)
+    .INPUT(x1, TensorType({DT_FLOAT16, DT_FLOAT, DT_DOUBLE, DT_UINT8, DT_INT8,
+                           DI_UINT16, DT_INT16, DT_INT32, DT_INT64,
+                           DT_COMPLEX64, DT_COMPLEX128}))
+    .INPUT(x2, TensorType({DT_FLOAT16, DT_FLOAT, DT_DOUBLE, DT_UINT8, DT_INT8,
+                           DI_UINT16, DT_INT16, DT_INT32, DT_INT64,
+                           DT_COMPLEX64, DT_COMPLEX128}))
+    .OUTPUT(y, TensorType({DT_FLOAT16, DT_FLOAT, DT_DOUBLE, DT_UINT8, DT_INT8,
+                           DI_UINT16, DT_INT16, DT_INT32, DT_INT64,
+                           DT_COMPLEX64, DT_COMPLEX128}))
+    .OP_END_FACTORY_REG(Mul)
 
 // for plugin
 static Status ParseParamsStub(const google::protobuf::Message* op_src, ge::Operator& op_dest) {
@@ -209,5 +386,36 @@ REGISTER_CUSTOM_OP("PartitionedCall")
     .OriginOpType({"ai.onnx::9::Clip"})
     .ParseParamsFn(ParseParamsClipV9Stub)
     .ParseOpToGraphFn(ParseOpToGraphClipV9Stub);
+REGISTER_CUSTOM_OP("TensorArray")
+    .FrameworkType(domi::TENSORFLOW)
+    .OriginOpType("TensorArrayV3")
+    .ParseParamsFn(ParseParamsStub);
+
+REGISTER_CUSTOM_OP("TensorArrayWrite")
+    .FrameworkType(domi::TENSORFLOW)
+    .OriginOpType("TensorArrayWriteV3")
+    .ParseParamsFn(ParseParamsStub);
+
+REGISTER_CUSTOM_OP("DynamicRNN")
+    .FrameworkType(domi::TENSORFLOW)
+    .OriginOpType("BlockLSTM")
+    .ParseParamsFn(ParseParamsStub);
+
+REGISTER_CUSTOM_OP("Merge")
+    .FrameworkType(domi::TENSORFLOW)
+    .OriginOpType("HistogramSummary")
+    .ParseParamsFn(ParseParamsStub);
+
+REGISTER_CUSTOM_OP("NoOp")
+    .FrameworkType(domi::TENSORFLOW)
+    .OriginOpType("NoOp")
+    .ParseParamsFn(ParseParamsStub);
+
+REGISTER_CUSTOM_OP("Fill")
+    .FrameworkType(domi::TENSORFLOW)
+    .OriginOpType("Fill")
+    .ParseParamsFn(ParseParamsStub);
 }  // namespace ge
+
+
 #endif  // MAIN_OPS_STUB_H

tests/st/testcase/origin_models/conv2d_depthwise_pb_gen.py

@@ -0,0 +1,26 @@
+import tensorflow as tf
+import os
+
+pb_file_path = os.getcwd()
+
+with tf.compat.v1.Session(graph=tf.Graph()) as sess:
+    # NHWC 
+    fmap_shape = [17, 101, 101, 17]
+    filter_size = [5, 5, 17, 1]
+    dy_shape = [17, 49, 49, 17]
+    strideh, stridew = [2, 2]
+    padding = 'VALID'
+    tensor_x1 = tf.compat.v1.placeholder(dtype="float16", shape=fmap_shape)
+    tensor_x2 = tf.compat.v1.placeholder(dtype="float16", shape=fmap_shape)
+    tensor_x = tf.add(tensor_x1, tensor_x2)
+    tensor_dy1 = tf.compat.v1.placeholder(dtype="float16", shape=dy_shape)
+    tensor_dy2 = tf.compat.v1.placeholder(dtype="float16", shape=dy_shape)
+    tensor_dy = tf.add(tensor_dy1, tensor_dy2)
+    
+    op = tf.nn.depthwise_conv2d_backprop_filter(tensor_x, filter_size, tensor_dy,
+                                  strides=[1, strideh, stridew, 1],
+                                  padding=padding,
+                                  data_format='NHWC',
+                                  dilations=[1,1,1,1])
+
+    tf.io.write_graph(sess.graph, logdir="./", name="test_depth_wise_conv2d.pb", as_text=False)

tests/st/testcase/origin_models/tensor_array_pb_gen.py

@@ -0,0 +1,14 @@
+import tensorflow as tf
+import os
+from tensorflow.python.framework import graph_util
+from tensorflow.python.ops import gen_data_flow_ops
+import numpy as np
+
+with tf.compat.v1.Session(graph=tf.Graph()) as sess:
+    size = tf.compat.v1.placeholder(dtype="int32", shape=())
+    value = tf.compat.v1.placeholder(dtype="float32", shape=(2,2))
+    index = tf.compat.v1.placeholder(dtype="int32", shape=())
+    flow = tf.compat.v1.placeholder(dtype="float32", shape=())
+    handleTensor = gen_data_flow_ops.tensor_array_v3(size= size, dtype = np.float32)
+    output = gen_data_flow_ops.tensor_array_write_v3(handle = handleTensor[0], index=index, value=value, flow_in=flow)
+    tf.io.write_graph(sess.graph, logdir="./", name="tensor_array.pb", as_text=False)

tests/st/testcase/origin_models/test.json

@@ -0,0 +1,71 @@
+{
+  "node": [
+    {
+      "attr": [
+        {
+          "key": "dtype",
+          "value": {
+            "type": "DT_HALF"
+          }
+        },
+        {
+          "key": "shape",
+          "value": {
+            "shape": {
+              "dim": [
+                {
+                  "size": 1
+                }
+              ]
+            }
+          }
+        }
+      ],
+      "name": "Placeholder",
+      "op": "Placeholder"
+    },
+    {
+      "attr": [
+        {
+          "key": "dtype",
+          "value": {
+            "type": "DT_HALF"
+          }
+        },
+        {
+          "key": "shape",
+          "value": {
+            "shape": {
+              "dim": [
+                {
+                  "size": 1
+                }
+              ]
+            }
+          }
+        }
+      ],
+      "name": "Placeholder_1",
+      "op": "Placeholder"
+    },
+    {
+      "attr": [
+        {
+          "key": "T",
+          "value": {
+            "type": "DT_HALF"
+          }
+        }
+      ],
+      "input": [
+        "Placeholder",
+        "Placeholder_1"
+      ],
+      "name": "add_test_1",
+      "op": "Add"
+    }
+  ],
+  "versions": {
+    "producer": 134
+  }
+}

tests/st/testcase/origin_models/test_VarIsInitializedOp_pb_gen.py

@@ -0,0 +1,14 @@
+import tensorflow as tf
+import numpy as np
+
+def generate_VarIsInitializedOp_pb():
+    with tf.compat.v1.Session(graph=tf.Graph()) as sess:
+        x = tf.compat.v1.placeholder(dtype="int32", shape=())
+        y = tf.Variable(tf.compat.v1.random_normal(shape=[4,3],mean=0,stddev=1), dtype="float32", name='y')
+        init = tf.compat.v1.global_variables_initializer()
+        sess.run(init)
+        op = tf.compat.v1.raw_ops.VarIsInitializedOp(resource=y, name="VarIsInitializedOp")
+    tf.io.write_graph(sess.graph, logdir="./", name="test_VarIsInitializedOp.pb", as_text=False)
+
+if __name__=='__main__':
+    generate_VarIsInitializedOp_pb()

tests/st/testcase/origin_models/test_avgpool3dgrad_pb_gen.py

@@ -0,0 +1,41 @@
+import tensorflow as tf
+import os
+import numpy as np
+from tensorflow.python.framework import graph_util
+
+pb_file_path = os.getcwd()
+
+def generate_case_1():
+    with tf.compat.v1.Session(graph=tf.Graph()) as sess:
+        grads_1 = tf.compat.v1.placeholder(dtype="float16", shape=(1,2,2,2,1))
+        grads_2 = tf.compat.v1.placeholder(dtype="float16", shape=(1,2,2,2,1))
+        grads = tf.add(grads_1, grads_2)
+        orig_input_shape = tf.constant(np.array([1,3,3,3,1]).astype("int32"))
+        op = tf.raw_ops.AvgPool3DGrad(orig_input_shape=orig_input_shape,
+                                      grad=grads,
+                                      ksize=[1,2,2,2,1],
+                                      strides=[1,1,1,1,1],
+                                      padding="VALID",
+                                      data_format='NDHWC',
+                                      name='AvgPool3DGrad')
+
+        tf.io.write_graph(sess.graph, logdir="./", name="avgpool3dgrad_case_1.pb", as_text=False)
+
+def generate_case_2():
+    with tf.compat.v1.Session(graph=tf.Graph()) as sess:
+        grads_1 = tf.compat.v1.placeholder(dtype="float16", shape=(1,1,2,2,2))
+        grads_2 = tf.compat.v1.placeholder(dtype="float16", shape=(1,1,2,2,2))
+        grads = tf.add(grads_1, grads_2)
+        orig_input_shape = tf.constant(np.array([1,1,3,3,3]).astype("int32"), )
+        op = tf.raw_ops.AvgPool3DGrad(orig_input_shape=orig_input_shape,
+                                      grad=grads,
+                                      ksize=[1,1,2,2,2],
+                                      strides=[1,1,1,1,1],
+                                      padding="VALID",
+                                      data_format='NCDHW',
+                                      name='AvgPool3DGrad')
+
+        tf.io.write_graph(sess.graph, logdir="./", name="avgpool3dgrad.pb.txt", as_text=False)
+
+if __name__=='__main__':
+    generate_case_2()

tests/st/testcase/origin_models/test_blocklstm_pb.gen.py

@@ -0,0 +1,34 @@
+import tensorflow as tf
+import os
+
+pb_file_path = os.getcwd()
+
+def generate_case_0():
+    with tf.compat.v1.Session(graph=tf.Graph()) as sess:
+        input_dtype = tf.float32
+        input_shape0 = [1, ]
+        input_shape1 = [202, 1, 768]
+        input_shape2 = [1, 1]
+        input_shape3 = [1, 1]
+        input_shape4 = [769, 4]
+        input_shape5 = [1, ]
+        input_shape6 = [1, ]
+        input_shape7 = [1, ]
+        input_shape8 = [4, ]
+
+        d0 = tf.compat.v1.placeholder(dtype=tf.int64, shape=input_shape0)
+        d1 = tf.compat.v1.placeholder(dtype=input_dtype, shape=input_shape1)
+        d2 = tf.compat.v1.placeholder(dtype=input_dtype, shape=input_shape2)
+        d3 = tf.compat.v1.placeholder(dtype=input_dtype, shape=input_shape3)
+        d4 = tf.compat.v1.placeholder(dtype=input_dtype, shape=input_shape4)
+        d5 = tf.compat.v1.placeholder(dtype=input_dtype, shape=input_shape5)
+        d6 = tf.compat.v1.placeholder(dtype=input_dtype, shape=input_shape6)
+        d7 = tf.compat.v1.placeholder(dtype=input_dtype, shape=input_shape7)
+        d8 = tf.compat.v1.placeholder(dtype=input_dtype, shape=input_shape8)
+
+        i1, cs1, f1, o1, ci1, co1, h1 = tf.raw_ops.BlockLSTM(seq_len_max=d0, x=d1, cs_prev=d2, h_prev=d3, w=d4, wci=d5, wcf=d6, wco=d7, b=d8,
+                                                             forget_bias=1, cell_clip=3, use_peephole=False, name="blockLSTM")
+        tf.io.write_graph(sess.graph, logdir="./", name="blocklstm_case.pb", as_text=False)
+
+if __name__=='__main__':
+    generate_case_0()

tests/st/testcase/origin_models/test_constant_pb_gen.py

@@ -0,0 +1,21 @@
+import tensorflow as tf
+import numpy as np
+from tensorflow.python.framework import graph_util
+
+def generate_constant_pb():
+    with tf.compat.v1.Session(graph=tf.Graph()) as sess:
+        grads_1 = tf.compat.v1.placeholder(dtype="float16", shape=(1,1,2,2,2))
+        grads_2 = tf.compat.v1.placeholder(dtype="float16", shape=(1,1,2,2,2))
+        grads = tf.add(grads_1, grads_2)
+        orig_input_shape = tf.constant(np.array([1,1,3,3,3]).astype("int32"), )
+        op = tf.raw_ops.AvgPool3DGrad(orig_input_shape=orig_input_shape,
+                                      grad=grads,
+                                      ksize=[1,1,2,2,2],
+                                      strides=[1,1,1,1,1],
+                                      padding="VALID",
+                                      data_format='NCDHW',
+                                      name='AvgPool3DGrad')
+        tf.io.write_graph(sess.graph, logdir="./", name="test_constant.pb", as_text=False)
+
+if __name__ == "__main__":
+    generate_constant_pb()

tests/st/testcase/origin_models/test_conv2d_pb_gen.py

@@ -0,0 +1,26 @@
+import tensorflow as tf
+import os
+from tensorflow.python.framework import graph_util
+
+pb_file_path = os.getcwd()
+
+def generate_conv2d_pb():
+    with tf.compat.v1.Session(graph=tf.Graph()) as sess:
+        input_x = tf.compat.v1.placeholder(dtype="float32", shape=(1,56,56,64))
+        input_filter = tf.compat.v1.placeholder(dtype="float32", shape=(3,3,64,64))
+        op = tf.nn.conv2d(input_x, input_filter, strides=[1,1,1,1], padding=[[0,0],[1,1],[1,1],[0,0]],
+                          data_format="NHWC", dilations=[1,1,1,1], name='conv2d_res')
+        tf.io.write_graph(sess.graph, logdir="./", name="conv2d.pb", as_text=False)
+
+def generate_add_pb():
+    with tf.compat.v1.Session(graph=tf.Graph()) as sess:
+        x = tf.compat.v1.placeholder(tf.int32, name='x')
+        y = tf.compat.v1.placeholder(tf.int32, name='y')
+        b = tf.Variable(1, name='b')
+        xy = tf.multiply(x, y)
+        op = tf.add(xy, b, name='op_to_store')
+        tf.io.write_graph(sess.graph, logdir="./", name="model.pb", as_text=False)
+
+if __name__=='__main__':
+    generate_conv2d_pb()
+    generate_add_pb()

tests/st/testcase/origin_models/test_enter_pb_gen.py

@@ -0,0 +1,13 @@
+import tensorflow as tf
+from tensorflow.python.ops import control_flow_ops
+
+def generate_enter_pb():
+    with tf.compat.v1.Session(graph=tf.Graph()) as sess:
+        x = tf.compat.v1.placeholder(dtype="int32", shape=())
+        y = tf.compat.v1.placeholder(dtype="int32", shape=())
+        output1 = control_flow_ops.enter(x, frame_name = "output1")
+        output2 = control_flow_ops.enter(y, frame_name = "output2")
+    tf.io.write_graph(sess.graph, logdir="./", name="test_enter.pb", as_text=False)
+
+if __name__=='__main__':
+    generate_enter_pb()

tests/st/testcase/origin_models/test_fill_pb_gen.py

@@ -0,0 +1,12 @@
+import tensorflow as tf
+import numpy as np
+
+def generate_fill_pb():
+    with tf.compat.v1.Session(graph=tf.Graph()) as sess:
+        x = tf.compat.v1.placeholder(dtype="int32", shape=(2,2))
+        y = tf.fill([1,2], value = 5)
+        z = tf.add(x,y)
+    tf.io.write_graph(sess.graph, logdir="./", name="test_fill.pb", as_text=False)
+
+if __name__ == "__main__":
+    generate_fill_pb()

tests/st/testcase/origin_models/test_identity_pb_gen.py

@@ -0,0 +1,13 @@
+import tensorflow as tf
+
+def generate_identity_pb():
+    with tf.compat.v1.Session(graph=tf.Graph()) as sess:
+        x = tf.compat.v1.placeholder(dtype="int32", shape=())
+        x_plus_1 = tf.add(x, 1, name='x_plus')
+        with tf.control_dependencies([x_plus_1]):
+            y = x
+            z = tf.identity(x,name='identity')
+    tf.io.write_graph(sess.graph, logdir="./", name="test_identity.pb", as_text=False)
+
+if __name__=='__main__':
+    generate_identity_pb()

tests/st/testcase/origin_models/test_merge_pb_gen.py

@@ -0,0 +1,18 @@
+import tensorflow as tf
+from tensorflow.python.framework import graph_util
+import numpy as np
+
+def generate_merge_pb():
+    with tf.compat.v1.Session(graph=tf.Graph()) as sess:
+        dist = tf.compat.v1.placeholder(tf.float32, [100])
+        tf.compat.v1.summary.histogram(name="Merge", values=dist)
+        writer = tf.compat.v1.summary.FileWriter("./tf_summary_merge_pb")
+        op = tf.compat.v1.summary.merge_all()
+        for step in range(10):
+            mean_moving_normal = np.random.normal(loc=step, scale=1, size=[100])
+            summ = sess.run(op, feed_dict = {dist : mean_moving_normal})
+            writer.add_summary(summ, global_step=step)
+        tf.io.write_graph(sess.graph, logdir="./", name="merge.pb", as_text=False)
+
+if __name__=='__main__':
+    generate_merge_pb()

tests/st/testcase/origin_models/test_no_op_pb_gen.py

@@ -0,0 +1,14 @@
+import tensorflow as tf
+import numpy as np
+
+def generate_no_op_pb():
+    with tf.compat.v1.Session(graph=tf.Graph()) as sess:
+        x = tf.compat.v1.placeholder(dtype="int32", shape=())
+        y = tf.compat.v1.placeholder(dtype="int32", shape=())
+        add_op = tf.add(x, y)
+        y2 = tf.no_op(name="train")
+        z = tf.group([add_op, y2])
+    tf.io.write_graph(sess.graph, logdir="./", name="test_no_op.pb", as_text=False)
+
+if __name__=='__main__':
+    generate_no_op_pb()

tests/st/testcase/origin_models/test_reshape_pb_gen.py

@@ -0,0 +1,10 @@
+import tensorflow as tf
+
+def generate_reshape_pb():
+    with tf.compat.v1.Session(graph=tf.Graph()) as sess:
+        x = tf.compat.v1.placeholder(dtype="int32", shape=(2,2))
+        y = tf.compat.v1.reshape(x, [2,2])
+    tf.io.write_graph(sess.graph, logdir="./", name="test_reshape.pb", as_text=False)
+
+if __name__ == "__main__":
+    generate_reshape_pb()

tests/st/testcase/origin_models/test_sequeeze_pb_gen.py

@@ -0,0 +1,13 @@
+import tensorflow as tf
+import numpy as np
+
+def generate_sequeeze_pb():
+    with tf.compat.v1.Session(graph=tf.Graph()) as sess:
+        x = tf.compat.v1.placeholder(dtype="int32", shape=(2,2))
+        y = tf.constant([[1,2],[2,3]])
+        z = tf.add(x,y)
+        op = tf.squeeze(z,name = "squeeze")
+    tf.io.write_graph(sess.graph, logdir="./", name="test_sequeeze.pb", as_text=False)
+
+if __name__ == "__main__":
+    generate_sequeeze_pb()

tests/st/testcase/origin_models/test_shape_n_pb_gen.py

@@ -0,0 +1,11 @@
+import tensorflow as tf
+import numpy as np
+
+def generate_shape_n_pb():
+    with tf.compat.v1.Session(graph=tf.Graph()) as sess:
+        x = tf.compat.v1.placeholder(dtype="int32", shape=(2,2))
+        y = tf.shape_n([1,2], name= "shape_n")
+    tf.io.write_graph(sess.graph, logdir="./", name="test_shape_n.pb", as_text=False)
+
+if __name__ == "__main__":
+    generate_shape_n_pb()

tests/st/testcase/origin_models/test_switch_pb_gen.py

@@ -0,0 +1,13 @@
+import tensorflow as tf
+from tensorflow.python.ops import control_flow_ops
+
+def generate_switch_pb():
+    with tf.compat.v1.Session(graph=tf.Graph()) as sess:
+        x = tf.compat.v1.placeholder(dtype="int32", shape=())
+        y = tf.compat.v1.placeholder(dtype="int32", shape=())
+        output1 = control_flow_ops.switch(x,False)
+        output2 = control_flow_ops.switch(y,True)
+    tf.io.write_graph(sess.graph, logdir="./", name="test_switch.pb", as_text=False)
+
+if __name__=='__main__':
+    generate_switch_pb()

tests/st/testcase/origin_models/test_variableV2_pb_gen.py

@@ -0,0 +1,13 @@
+import tensorflow as tf
+
+def generate_VariableV2_pb():
+    with tf.compat.v1.Session(graph=tf.Graph()) as sess:
+        x = tf.compat.v1.placeholder(dtype="int32", shape=(2,3))
+        op = tf.raw_ops.VariableV2(shape=[2,3], dtype="int32", name="VariableV2")
+        init = tf.compat.v1.global_variables_initializer()
+        op_add = tf.add(x,op)
+        sess.run(init)
+    tf.io.write_graph(sess.graph, logdir="./", name="test_VariableV2.pb", as_text=False)
+
+if __name__=='__main__':
+    generate_VariableV2_pb()