convert pnnx torch.unbind and torch.ones/ones_like family (#3583)

* convert torch.unbind

* torch.ones torch.ones_like

* torch.full torch.full_like

* torch.randn_like

* torch.empty torch.empty_like

tools/pnnx/src/CMakeLists.txt

@@ -185,16 +185,23 @@ set(pnnx_pass_level2_SRCS
     pass_level2/torch_clamp.cpp
     pass_level2/torch_clone.cpp
     pass_level2/torch_dequantize.cpp
+    pass_level2/torch_empty.cpp
+    pass_level2/torch_empty_like.cpp
     pass_level2/torch_flatten.cpp
     pass_level2/torch_flip.cpp
+    pass_level2/torch_full.cpp
+    pass_level2/torch_full_like.cpp
     pass_level2/torch_logsumexp.cpp
     pass_level2/torch_matmul.cpp
     pass_level2/torch_mean.cpp
     pass_level2/torch_norm.cpp
     pass_level2/torch_normal.cpp
+    pass_level2/torch_ones.cpp
+    pass_level2/torch_ones_like.cpp
     pass_level2/torch_prod.cpp
     pass_level2/torch_quantize_per_tensor.cpp
     pass_level2/torch_randn.cpp
+    pass_level2/torch_randn_like.cpp
     pass_level2/torch_roll.cpp
     pass_level2/torch_split.cpp
     pass_level2/torch_squeeze.cpp
@@ -263,6 +270,7 @@ set(pnnx_pass_ncnn_SRCS
     pass_ncnn/convert_torch_cat.cpp
     pass_ncnn/convert_torch_chunk.cpp
     pass_ncnn/convert_torch_split.cpp
+    pass_ncnn/convert_torch_unbind.cpp
     pass_ncnn/eliminate_output.cpp
     pass_ncnn/expand_expression.cpp
     pass_ncnn/insert_split.cpp

tools/pnnx/src/ir.cpp

@@ -1735,6 +1735,10 @@ int Graph::python(const std::string& pypath, const std::string& pnnxbinpath)
                     {
                         fprintf(pyfp, "%s=", it.first.c_str());
                     }
+                    else if (op->inputs.empty() && i == 0)
+                    {
+                        fprintf(pyfp, "%s=", it.first.c_str());
+                    }
                     else
                     {
                         fprintf(pyfp, ", %s=", it.first.c_str());

tools/pnnx/src/pass_level2/torch_empty.cpp

@@ -0,0 +1,45 @@
+// Tencent is pleased to support the open source community by making ncnn available.
+//
+// Copyright (C) 2022 THL A29 Limited, a Tencent company. All rights reserved.
+//
+// Licensed under the BSD 3-Clause License (the "License"); you may not use this file except
+// in compliance with the License. You may obtain a copy of the License at
+//
+// https://opensource.org/licenses/BSD-3-Clause
+//
+// 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.
+
+#include "pass_level2.h"
+
+namespace pnnx {
+
+class torch_empty : public GraphRewriterPass
+{
+public:
+    const char* match_pattern_graph() const
+    {
+        return R"PNNXIR(7767517
+8 7
+pnnx.Input              input_0     0 1 size
+prim::Constant          op_0        0 1 dtype value=*
+prim::Constant          op_1        0 1 layout value=*
+prim::Constant          op_2        0 1 device value=*
+prim::Constant          op_3        0 1 requires_grad value=*
+prim::Constant          op_4        0 1 memory_format value=*
+aten::empty             op_5        6 1 size dtype layout device requires_grad memory_format out
+pnnx.Output             output      1 0 out
+)PNNXIR";
+    }
+
+    const char* type_str() const
+    {
+        return "torch.empty";
+    }
+};
+
+REGISTER_GLOBAL_PNNX_GRAPH_REWRITER_PASS(torch_empty, 20)
+
+} // namespace pnnx

tools/pnnx/src/pass_level2/torch_empty_like.cpp

@@ -0,0 +1,45 @@
+// Tencent is pleased to support the open source community by making ncnn available.
+//
+// Copyright (C) 2022 THL A29 Limited, a Tencent company. All rights reserved.
+//
+// Licensed under the BSD 3-Clause License (the "License"); you may not use this file except
+// in compliance with the License. You may obtain a copy of the License at
+//
+// https://opensource.org/licenses/BSD-3-Clause
+//
+// 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.
+
+#include "pass_level2.h"
+
+namespace pnnx {
+
+class torch_empty_like : public GraphRewriterPass
+{
+public:
+    const char* match_pattern_graph() const
+    {
+        return R"PNNXIR(7767517
+8 7
+pnnx.Input              input_0     0 1 input
+prim::Constant          op_0        0 1 dtype value=*
+prim::Constant          op_1        0 1 layout value=*
+prim::Constant          op_2        0 1 device value=*
+prim::Constant          op_3        0 1 requires_grad value=*
+prim::Constant          op_4        0 1 memory_format value=*
+aten::empty_like        op_5        6 1 input dtype layout device requires_grad memory_format out
+pnnx.Output             output      1 0 out
+)PNNXIR";
+    }
+
+    const char* type_str() const
+    {
+        return "torch.empty_like";
+    }
+};
+
+REGISTER_GLOBAL_PNNX_GRAPH_REWRITER_PASS(torch_empty_like, 20)
+
+} // namespace pnnx

tools/pnnx/src/pass_level2/torch_full.cpp

@@ -0,0 +1,45 @@
+// Tencent is pleased to support the open source community by making ncnn available.
+//
+// Copyright (C) 2022 THL A29 Limited, a Tencent company. All rights reserved.
+//
+// Licensed under the BSD 3-Clause License (the "License"); you may not use this file except
+// in compliance with the License. You may obtain a copy of the License at
+//
+// https://opensource.org/licenses/BSD-3-Clause
+//
+// 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.
+
+#include "pass_level2.h"
+
+namespace pnnx {
+
+class torch_full : public GraphRewriterPass
+{
+public:
+    const char* match_pattern_graph() const
+    {
+        return R"PNNXIR(7767517
+8 7
+pnnx.Input              input_0     0 1 size
+pnnx.Input              input_1     0 1 fill_value
+prim::Constant          op_0        0 1 dtype value=*
+prim::Constant          op_1        0 1 layout value=*
+prim::Constant          op_2        0 1 device value=*
+prim::Constant          op_3        0 1 requires_grad value=*
+aten::full              op_4        6 1 size fill_value dtype layout device requires_grad out
+pnnx.Output             output      1 0 out
+)PNNXIR";
+    }
+
+    const char* type_str() const
+    {
+        return "torch.full";
+    }
+};
+
+REGISTER_GLOBAL_PNNX_GRAPH_REWRITER_PASS(torch_full, 20)
+
+} // namespace pnnx

tools/pnnx/src/pass_level2/torch_full_like.cpp

@@ -0,0 +1,46 @@
+// Tencent is pleased to support the open source community by making ncnn available.
+//
+// Copyright (C) 2021 THL A29 Limited, a Tencent company. All rights reserved.
+//
+// Licensed under the BSD 3-Clause License (the "License"); you may not use this file except
+// in compliance with the License. You may obtain a copy of the License at
+//
+// https://opensource.org/licenses/BSD-3-Clause
+//
+// 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.
+
+#include "pass_level2.h"
+
+namespace pnnx {
+
+class torch_full_like : public GraphRewriterPass
+{
+public:
+    const char* match_pattern_graph() const
+    {
+        return R"PNNXIR(7767517
+9 8
+pnnx.Input              input_0     0 1 input
+pnnx.Input              input_1     0 1 fill_value
+prim::Constant          op_0        0 1 dtype value=*
+prim::Constant          op_1        0 1 layout value=*
+prim::Constant          op_2        0 1 device value=*
+prim::Constant          op_3        0 1 requires_grad value=*
+prim::Constant          op_4        0 1 memory_format value=*
+aten::full_like         op_5        7 1 input fill_value dtype layout device requires_grad memory_format out
+pnnx.Output             output      1 0 out
+)PNNXIR";
+    }
+
+    const char* type_str() const
+    {
+        return "torch.full_like";
+    }
+};
+
+REGISTER_GLOBAL_PNNX_GRAPH_REWRITER_PASS(torch_full_like, 20)
+
+} // namespace pnnx

tools/pnnx/src/pass_level2/torch_ones.cpp

@@ -0,0 +1,44 @@
+// Tencent is pleased to support the open source community by making ncnn available.
+//
+// Copyright (C) 2022 THL A29 Limited, a Tencent company. All rights reserved.
+//
+// Licensed under the BSD 3-Clause License (the "License"); you may not use this file except
+// in compliance with the License. You may obtain a copy of the License at
+//
+// https://opensource.org/licenses/BSD-3-Clause
+//
+// 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.
+
+#include "pass_level2.h"
+
+namespace pnnx {
+
+class torch_ones : public GraphRewriterPass
+{
+public:
+    const char* match_pattern_graph() const
+    {
+        return R"PNNXIR(7767517
+7 6
+pnnx.Input              input_0     0 1 size
+prim::Constant          op_0        0 1 dtype value=*
+prim::Constant          op_1        0 1 layout value=*
+prim::Constant          op_2        0 1 device value=*
+prim::Constant          op_3        0 1 requires_grad value=*
+aten::ones              op_4        5 1 size dtype layout device requires_grad out
+pnnx.Output             output      1 0 out
+)PNNXIR";
+    }
+
+    const char* type_str() const
+    {
+        return "torch.ones";
+    }
+};
+
+REGISTER_GLOBAL_PNNX_GRAPH_REWRITER_PASS(torch_ones, 20)
+
+} // namespace pnnx

tools/pnnx/src/pass_level2/torch_ones_like.cpp

@@ -0,0 +1,45 @@
+// Tencent is pleased to support the open source community by making ncnn available.
+//
+// Copyright (C) 2021 THL A29 Limited, a Tencent company. All rights reserved.
+//
+// Licensed under the BSD 3-Clause License (the "License"); you may not use this file except
+// in compliance with the License. You may obtain a copy of the License at
+//
+// https://opensource.org/licenses/BSD-3-Clause
+//
+// 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.
+
+#include "pass_level2.h"
+
+namespace pnnx {
+
+class torch_ones_like : public GraphRewriterPass
+{
+public:
+    const char* match_pattern_graph() const
+    {
+        return R"PNNXIR(7767517
+8 7
+pnnx.Input              input_0     0 1 input
+prim::Constant          op_0        0 1 dtype value=*
+prim::Constant          op_1        0 1 layout value=*
+prim::Constant          op_2        0 1 device value=*
+prim::Constant          op_3        0 1 requires_grad value=*
+prim::Constant          op_4        0 1 memory_format value=*
+aten::ones_like         op_5        6 1 input dtype layout device requires_grad memory_format out
+pnnx.Output             output      1 0 out
+)PNNXIR";
+    }
+
+    const char* type_str() const
+    {
+        return "torch.ones_like";
+    }
+};
+
+REGISTER_GLOBAL_PNNX_GRAPH_REWRITER_PASS(torch_ones_like, 20)
+
+} // namespace pnnx

tools/pnnx/src/pass_level2/torch_randn_like.cpp

@@ -0,0 +1,45 @@
+// Tencent is pleased to support the open source community by making ncnn available.
+//
+// Copyright (C) 2022 THL A29 Limited, a Tencent company. All rights reserved.
+//
+// Licensed under the BSD 3-Clause License (the "License"); you may not use this file except
+// in compliance with the License. You may obtain a copy of the License at
+//
+// https://opensource.org/licenses/BSD-3-Clause
+//
+// 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.
+
+#include "pass_level2.h"
+
+namespace pnnx {
+
+class torch_randn_like : public GraphRewriterPass
+{
+public:
+    const char* match_pattern_graph() const
+    {
+        return R"PNNXIR(7767517
+8 7
+pnnx.Input              input_0     0 1 input
+prim::Constant          op_0        0 1 dtype value=*
+prim::Constant          op_1        0 1 layout value=*
+prim::Constant          op_2        0 1 device value=*
+prim::Constant          op_3        0 1 requires_grad value=*
+prim::Constant          op_4        0 1 memory_format value=*
+aten::randn_like        op_5        6 1 input dtype layout device requires_grad memory_format out
+pnnx.Output             output      1 0 out
+)PNNXIR";
+    }
+
+    const char* type_str() const
+    {
+        return "torch.randn_like";
+    }
+};
+
+REGISTER_GLOBAL_PNNX_GRAPH_REWRITER_PASS(torch_randn_like, 20)
+
+} // namespace pnnx

tools/pnnx/src/pass_ncnn.cpp

@@ -21,6 +21,7 @@
 #include "pass_ncnn/convert_torch_cat.h"
 #include "pass_ncnn/convert_torch_chunk.h"
 #include "pass_ncnn/convert_torch_split.h"
+#include "pass_ncnn/convert_torch_unbind.h"
 #include "pass_ncnn/eliminate_output.h"
 #include "pass_ncnn/expand_expression.h"
 #include "pass_ncnn/insert_split.h"
@@ -80,6 +81,11 @@ void pass_ncnn(Graph& g)
 
     ncnn::convert_half_to_float(g);
 
+    ncnn::convert_torch_cat(g);
+    ncnn::convert_torch_chunk(g);
+    ncnn::convert_torch_split(g);
+    ncnn::convert_torch_unbind(g);
+
     int opindex = 0;
     for (auto x : g_global_pnnx_ncnn_graph_rewriter_passes)
     {
@@ -89,10 +95,6 @@ void pass_ncnn(Graph& g)
         }
     }
 
-    ncnn::convert_torch_cat(g);
-    ncnn::convert_torch_chunk(g);
-    ncnn::convert_torch_split(g);
-
     ncnn::insert_split(g);
 
     ncnn::eliminate_noop(g);

tools/pnnx/src/pass_ncnn/convert_torch_unbind.cpp

@@ -0,0 +1,97 @@
+// Tencent is pleased to support the open source community by making ncnn available.
+//
+// Copyright (C) 2022 THL A29 Limited, a Tencent company. All rights reserved.
+//
+// Licensed under the BSD 3-Clause License (the "License"); you may not use this file except
+// in compliance with the License. You may obtain a copy of the License at
+//
+// https://opensource.org/licenses/BSD-3-Clause
+//
+// 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.
+
+#include "convert_torch_unbind.h"
+
+namespace pnnx {
+
+namespace ncnn {
+
+void convert_torch_unbind(Graph& graph)
+{
+    int op_index = 0;
+
+    while (1)
+    {
+        bool matched = false;
+
+        for (Operator* op : graph.ops)
+        {
+            if (op->type != "torch.unbind")
+                continue;
+
+            matched = true;
+
+            op->type = "Slice";
+            op->name = std::string("unbind_") + std::to_string(op_index++);
+
+            const int batch_index = op->inputs[0]->params["__batch_index"].i;
+
+            int axis = op->params.at("dim").i;
+            if (axis == batch_index)
+            {
+                fprintf(stderr, "unbind along batch axis %d is not supported\n", batch_index);
+                continue;
+            }
+
+            if (axis < 0)
+            {
+                int input_rank = op->inputs[0]->shape.size();
+                axis = input_rank + axis;
+            }
+
+            int output_size = (int)op->outputs.size();
+
+            if (axis > batch_index)
+                axis -= 1;
+
+            op->params["0"].type = 5;
+            op->params["0"].ai.resize(output_size, -233);
+
+            op->params["1"] = axis;
+
+            op->params.erase("dim");
+
+            // reshape for each output, squeezing the unbind dim
+            for (int i = 0; i < output_size; i++)
+            {
+                Operand* out = op->outputs[i];
+
+                Operator* reshape = graph.new_operator_after("Tensor.reshape", op->name + "_ncnnreshape" + std::to_string(i), op);
+
+                Operand* reshape_in = graph.new_operand(op->name + "_ncnnreshape" + std::to_string(i) + "_in");
+
+                reshape->inputs.push_back(reshape_in);
+                reshape->outputs.push_back(out);
+
+                op->outputs[i] = reshape_in;
+
+                out->producer = reshape;
+                reshape_in->producer = op;
+                reshape_in->consumers.push_back(reshape);
+
+                reshape->params["shape"] = out->shape;
+            }
+
+            break;
+        }
+
+        if (!matched)
+            break;
+    }
+}
+
+} // namespace ncnn
+
+} // namespace pnnx

tools/pnnx/src/pass_ncnn/convert_torch_unbind.h

@@ -0,0 +1,25 @@
+// Tencent is pleased to support the open source community by making ncnn available.
+//
+// Copyright (C) 2022 THL A29 Limited, a Tencent company. All rights reserved.
+//
+// Licensed under the BSD 3-Clause License (the "License"); you may not use this file except
+// in compliance with the License. You may obtain a copy of the License at
+//
+// https://opensource.org/licenses/BSD-3-Clause
+//
+// 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.
+
+#include "pass_ncnn.h"
+
+namespace pnnx {
+
+namespace ncnn {
+
+void convert_torch_unbind(Graph& graph);
+
+} // namespace ncnn
+
+} // namespace pnnx

tools/pnnx/tests/CMakeLists.txt

@@ -171,10 +171,14 @@ pnnx_add_test(torch_chunk)
 pnnx_add_test(torch_clamp)
 pnnx_add_test(torch_clone)
 pnnx_add_test(torch_flatten)
+pnnx_add_test(torch_full)
+pnnx_add_test(torch_full_like)
 pnnx_add_test(torch_logsumexp)
 pnnx_add_test(torch_matmul)
 pnnx_add_test(torch_mean)
 pnnx_add_test(torch_norm)
+pnnx_add_test(torch_ones)
+pnnx_add_test(torch_ones_like)
 pnnx_add_test(torch_permute)
 pnnx_add_test(torch_prod)
 pnnx_add_test(torch_sum)
@@ -184,6 +188,8 @@ pnnx_add_test(torch_stack)
 pnnx_add_test(torch_transpose)
 pnnx_add_test(torch_unbind)
 pnnx_add_test(torch_unsqueeze)
+pnnx_add_test(torch_zeros)
+pnnx_add_test(torch_zeros_like)
 
 pnnx_add_test(mobilenet_v2)
 pnnx_add_test(mobilenet_v3_small)

tools/pnnx/tests/ncnn/CMakeLists.txt

@@ -139,6 +139,7 @@ pnnx_ncnn_add_test(torch_prod)
 pnnx_ncnn_add_test(torch_sum)
 pnnx_ncnn_add_test(torch_squeeze)
 pnnx_ncnn_add_test(torch_transpose)
+pnnx_ncnn_add_test(torch_unbind)
 pnnx_ncnn_add_test(torch_unsqueeze)
 
 pnnx_ncnn_add_test(mobilenet_v2)

tools/pnnx/tests/ncnn/test_torch_unbind.py

@@ -0,0 +1,71 @@
+# Tencent is pleased to support the open source community by making ncnn available.
+#
+# Copyright (C) 2022 THL A29 Limited, a Tencent company. All rights reserved.
+#
+# Licensed under the BSD 3-Clause License (the "License"); you may not use this file except
+# in compliance with the License. You may obtain a copy of the License at
+#
+# https://opensource.org/licenses/BSD-3-Clause
+#
+# 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 torch
+import torch.nn as nn
+import torch.nn.functional as F
+
+class Model(nn.Module):
+    def __init__(self):
+        super(Model, self).__init__()
+
+    def forward(self, x, y):
+        x0, x1, x2 = torch.unbind(x, dim=0)
+        y0, y1, y2, y3, y4, y5, y6, y7, y8 = torch.unbind(y, dim=1)
+
+        x0 = F.relu(x0)
+        x1 = F.relu(x1)
+        y0 = F.relu(y0)
+        y1 = F.relu(y1)
+        y2 = F.relu(y2)
+        y3 = F.relu(y3)
+        y4 = F.relu(y4)
+        y5 = F.relu(y5)
+        y6 = F.relu(y6)
+        y7 = F.relu(y7)
+        y8 = F.relu(y8)
+        return x0, x1, y0, y1, y2, y3, y4, y5, y6, y7, y8
+
+def test():
+    net = Model()
+    net.eval()
+
+    torch.manual_seed(0)
+    x = torch.rand(3, 16)
+    y = torch.rand(5, 9, 11)
+
+    a = net(x, y)
+
+    # export torchscript
+    mod = torch.jit.trace(net, (x, y))
+    mod.save("test_torch_unbind.pt")
+
+    # torchscript to pnnx
+    import os
+    os.system("../../src/pnnx test_torch_unbind.pt inputshape=[3,16],[5,9,11]")
+
+    # ncnn inference
+    import test_torch_unbind_ncnn
+    b = test_torch_unbind_ncnn.test_inference()
+
+    for a0, b0 in zip(a, b):
+        if not torch.equal(a0, b0):
+            return False
+    return True
+
+if __name__ == "__main__":
+    if test():
+        exit(0)
+    else:
+        exit(1)

tools/pnnx/tests/test_torch_full.py

@@ -0,0 +1,61 @@
+# Tencent is pleased to support the open source community by making ncnn available.
+#
+# Copyright (C) 2022 THL A29 Limited, a Tencent company. All rights reserved.
+#
+# Licensed under the BSD 3-Clause License (the "License"); you may not use this file except
+# in compliance with the License. You may obtain a copy of the License at
+#
+# https://opensource.org/licenses/BSD-3-Clause
+#
+# 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 torch
+import torch.nn as nn
+import torch.nn.functional as F
+
+class Model(nn.Module):
+    def __init__(self):
+        super(Model, self).__init__()
+
+    def forward(self, x, y, z):
+        x = torch.full(x.size(), 1.5)
+        y = torch.full(y.size(), 3)
+        z = torch.full(z.size(), -2.2)
+        return x, y, z
+
+def test():
+    net = Model()
+    net.eval()
+
+    torch.manual_seed(0)
+    x = torch.rand(1, 3, 16)
+    y = torch.rand(1, 5, 9, 11)
+    z = torch.rand(14, 8, 5, 9, 10)
+
+    a = net(x, y, z)
+
+    # export torchscript
+    mod = torch.jit.trace(net, (x, y, z))
+    mod.save("test_torch_full.pt")
+
+    # torchscript to pnnx
+    import os
+    os.system("../src/pnnx test_torch_full.pt inputshape=[1,3,16],[1,5,9,11],[14,8,5,9,10]")
+
+    # pnnx inference
+    import test_torch_full_pnnx
+    b = test_torch_full_pnnx.test_inference()
+
+    for a0, b0 in zip(a, b):
+        if not torch.equal(a0, b0):
+            return False
+    return True
+
+if __name__ == "__main__":
+    if test():
+        exit(0)
+    else:
+        exit(1)

tools/pnnx/tests/test_torch_full_like.py

@@ -0,0 +1,61 @@
+# Tencent is pleased to support the open source community by making ncnn available.
+#
+# Copyright (C) 2022 THL A29 Limited, a Tencent company. All rights reserved.
+#
+# Licensed under the BSD 3-Clause License (the "License"); you may not use this file except
+# in compliance with the License. You may obtain a copy of the License at
+#
+# https://opensource.org/licenses/BSD-3-Clause
+#
+# 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 torch
+import torch.nn as nn
+import torch.nn.functional as F
+
+class Model(nn.Module):
+    def __init__(self):
+        super(Model, self).__init__()
+
+    def forward(self, x, y, z):
+        x = torch.full_like(x, 1.5)
+        y = torch.full_like(y, 3)
+        z = torch.full_like(z, -2.2)
+        return x, y, z
+
+def test():
+    net = Model()
+    net.eval()
+
+    torch.manual_seed(0)
+    x = torch.rand(1, 3, 16)
+    y = torch.rand(1, 5, 9, 11)
+    z = torch.rand(14, 8, 5, 9, 10)
+
+    a = net(x, y, z)
+
+    # export torchscript
+    mod = torch.jit.trace(net, (x, y, z))
+    mod.save("test_torch_full_like.pt")
+
+    # torchscript to pnnx
+    import os
+    os.system("../src/pnnx test_torch_full_like.pt inputshape=[1,3,16],[1,5,9,11],[14,8,5,9,10]")
+
+    # pnnx inference
+    import test_torch_full_like_pnnx
+    b = test_torch_full_like_pnnx.test_inference()
+
+    for a0, b0 in zip(a, b):
+        if not torch.equal(a0, b0):
+            return False
+    return True
+
+if __name__ == "__main__":
+    if test():
+        exit(0)
+    else:
+        exit(1)

tools/pnnx/tests/test_torch_ones.py

@@ -0,0 +1,61 @@
+# Tencent is pleased to support the open source community by making ncnn available.
+#
+# Copyright (C) 2022 THL A29 Limited, a Tencent company. All rights reserved.
+#
+# Licensed under the BSD 3-Clause License (the "License"); you may not use this file except
+# in compliance with the License. You may obtain a copy of the License at
+#
+# https://opensource.org/licenses/BSD-3-Clause
+#
+# 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 torch
+import torch.nn as nn
+import torch.nn.functional as F
+
+class Model(nn.Module):
+    def __init__(self):
+        super(Model, self).__init__()
+
+    def forward(self, x, y, z):
+        x = torch.ones(x.size())
+        y = torch.ones(y.size())
+        z = torch.ones(z.size())
+        return x, y, z
+
+def test():
+    net = Model()
+    net.eval()
+
+    torch.manual_seed(0)
+    x = torch.rand(1, 3, 16)
+    y = torch.rand(1, 5, 9, 11)
+    z = torch.rand(14, 8, 5, 9, 10)
+
+    a = net(x, y, z)
+
+    # export torchscript
+    mod = torch.jit.trace(net, (x, y, z))
+    mod.save("test_torch_ones.pt")
+
+    # torchscript to pnnx
+    import os
+    os.system("../src/pnnx test_torch_ones.pt inputshape=[1,3,16],[1,5,9,11],[14,8,5,9,10]")
+
+    # pnnx inference
+    import test_torch_ones_pnnx
+    b = test_torch_ones_pnnx.test_inference()
+
+    for a0, b0 in zip(a, b):
+        if not torch.equal(a0, b0):
+            return False
+    return True
+
+if __name__ == "__main__":
+    if test():
+        exit(0)
+    else:
+        exit(1)

tools/pnnx/tests/test_torch_ones_like.py

@@ -0,0 +1,61 @@
+# Tencent is pleased to support the open source community by making ncnn available.
+#
+# Copyright (C) 2022 THL A29 Limited, a Tencent company. All rights reserved.
+#
+# Licensed under the BSD 3-Clause License (the "License"); you may not use this file except
+# in compliance with the License. You may obtain a copy of the License at
+#
+# https://opensource.org/licenses/BSD-3-Clause
+#
+# 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 torch
+import torch.nn as nn
+import torch.nn.functional as F
+
+class Model(nn.Module):
+    def __init__(self):
+        super(Model, self).__init__()
+
+    def forward(self, x, y, z):
+        x = torch.ones_like(x)
+        y = torch.ones_like(y)
+        z = torch.ones_like(z)
+        return x, y, z
+
+def test():
+    net = Model()
+    net.eval()
+
+    torch.manual_seed(0)
+    x = torch.rand(1, 3, 16)
+    y = torch.rand(1, 5, 9, 11)
+    z = torch.rand(14, 8, 5, 9, 10)
+
+    a = net(x, y, z)
+
+    # export torchscript
+    mod = torch.jit.trace(net, (x, y, z))
+    mod.save("test_torch_ones_like.pt")
+
+    # torchscript to pnnx
+    import os
+    os.system("../src/pnnx test_torch_ones_like.pt inputshape=[1,3,16],[1,5,9,11],[14,8,5,9,10]")
+
+    # pnnx inference
+    import test_torch_ones_like_pnnx
+    b = test_torch_ones_like_pnnx.test_inference()
+
+    for a0, b0 in zip(a, b):
+        if not torch.equal(a0, b0):
+            return False
+    return True
+
+if __name__ == "__main__":
+    if test():
+        exit(0)
+    else:
+        exit(1)

tools/pnnx/tests/test_torch_zeros.py

@@ -0,0 +1,61 @@
+# Tencent is pleased to support the open source community by making ncnn available.
+#
+# Copyright (C) 2022 THL A29 Limited, a Tencent company. All rights reserved.
+#
+# Licensed under the BSD 3-Clause License (the "License"); you may not use this file except
+# in compliance with the License. You may obtain a copy of the License at
+#
+# https://opensource.org/licenses/BSD-3-Clause
+#
+# 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 torch
+import torch.nn as nn
+import torch.nn.functional as F
+
+class Model(nn.Module):
+    def __init__(self):
+        super(Model, self).__init__()
+
+    def forward(self, x, y, z):
+        x = torch.zeros(x.size())
+        y = torch.zeros(y.size())
+        z = torch.zeros(z.size())
+        return x, y, z
+
+def test():
+    net = Model()
+    net.eval()
+
+    torch.manual_seed(0)
+    x = torch.rand(1, 3, 16)
+    y = torch.rand(1, 5, 9, 11)
+    z = torch.rand(14, 8, 5, 9, 10)
+
+    a = net(x, y, z)
+
+    # export torchscript
+    mod = torch.jit.trace(net, (x, y, z))
+    mod.save("test_torch_zeros.pt")
+
+    # torchscript to pnnx
+    import os
+    os.system("../src/pnnx test_torch_zeros.pt inputshape=[1,3,16],[1,5,9,11],[14,8,5,9,10]")
+
+    # pnnx inference
+    import test_torch_zeros_pnnx
+    b = test_torch_zeros_pnnx.test_inference()
+
+    for a0, b0 in zip(a, b):
+        if not torch.equal(a0, b0):
+            return False
+    return True
+
+if __name__ == "__main__":
+    if test():
+        exit(0)
+    else:
+        exit(1)

tools/pnnx/tests/test_torch_zeros_like.py

@@ -0,0 +1,61 @@
+# Tencent is pleased to support the open source community by making ncnn available.
+#
+# Copyright (C) 2022 THL A29 Limited, a Tencent company. All rights reserved.
+#
+# Licensed under the BSD 3-Clause License (the "License"); you may not use this file except
+# in compliance with the License. You may obtain a copy of the License at
+#
+# https://opensource.org/licenses/BSD-3-Clause
+#
+# 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 torch
+import torch.nn as nn
+import torch.nn.functional as F
+
+class Model(nn.Module):
+    def __init__(self):
+        super(Model, self).__init__()
+
+    def forward(self, x, y, z):
+        x = torch.zeros_like(x)
+        y = torch.zeros_like(y)
+        z = torch.zeros_like(z)
+        return x, y, z
+
+def test():
+    net = Model()
+    net.eval()
+
+    torch.manual_seed(0)
+    x = torch.rand(1, 3, 16)
+    y = torch.rand(1, 5, 9, 11)
+    z = torch.rand(14, 8, 5, 9, 10)
+
+    a = net(x, y, z)
+
+    # export torchscript
+    mod = torch.jit.trace(net, (x, y, z))
+    mod.save("test_torch_zeros_like.pt")
+
+    # torchscript to pnnx
+    import os
+    os.system("../src/pnnx test_torch_zeros_like.pt inputshape=[1,3,16],[1,5,9,11],[14,8,5,9,10]")
+
+    # pnnx inference
+    import test_torch_zeros_like_pnnx
+    b = test_torch_zeros_like_pnnx.test_inference()
+
+    for a0, b0 in zip(a, b):
+        if not torch.equal(a0, b0):
+            return False
+    return True
+
+if __name__ == "__main__":
+    if test():
+        exit(0)
+    else:
+        exit(1)