feat: 前端支持 concat 及单元测试

Signed-off-by: YdrMaster <ydrml@hotmail.com>

include/core/graph_handler.h

@@ -58,6 +58,7 @@ class GraphHandlerObj {
     Tensor identity(Tensor x, Tensor y);
     Tensor flatten(Tensor s, Tensor y);
     Tensor reshape(Tensor data, Tensor reshaped, Shape shape);
+    Tensor concat(TensorVec inputs, Tensor output, int dim);
 };
 
 } // namespace infini

pyinfinitensor/src/pyinfinitensor/onnx.py

@@ -1,4 +1,5 @@
 import onnx, backend
+from typing import Dict
 
 runtime = backend.cpu_runtime()
 
@@ -6,8 +7,8 @@ runtime = backend.cpu_runtime()
 def from_onnx(model: onnx.ModelProto):
     handler = backend.GraphHandlerObj(runtime)
 
-    tensors = dict()
+    tensors: Dict[str, backend.TensorObj] = dict()
-    data = dict()
+    data: Dict[str, onnx.TensorProto] = dict()
 
     for input in model.graph.input:
         dims = [d.dim_value for d in input.type.tensor_type.shape.dim]
@@ -121,6 +122,12 @@ def from_onnx(model: onnx.ModelProto):
                 tensors.get(node.output[0]),
                 [int(i) for i in data[node.input[1]].int64_data],
             )
+        elif node.op_type == "Concat":
+            tensors[node.output[0]] = handler.concat(
+                [tensors[name] for name in node.input],
+                tensors.get(node.output[0]),
+                next((attr.i for attr in node.attribute if attr.name == "axis")),
+            )
         else:
             raise Exception('Unsupported operator "{}"'.format(node.op_type))
 

pyinfinitensor/tests/test_onnx.py

@@ -136,11 +136,11 @@ class TestStringMethods(unittest.TestCase):
         make_and_import_model(make_graph([flatten], "flatten", [x], [y]))
 
     def test_reshape(self):
-        data = make_tensor_value_info("data", TensorProto.FLOAT, [2, 3, 3, 4])
+        data = make_tensor_value_info("data", TensorProto.FLOAT, [2, 3, 4, 5])
         # shape 对于后端来说并不是一个张量，然而转换中可能没有办法分辨
         # 不知道怎么把 ValueInfoProto 转换成 TensorProto
-        shape = make_tensor_value_info("shape", TensorProto.INT64, [4])
+        shape = make_tensor_value_info("shape", TensorProto.INT64, [3])
-        shape_data = make_tensor("shape", TensorProto.INT64, [4], [3, 2, 4, 3])
+        shape_data = make_tensor("shape", TensorProto.INT64, [3], [5, 3, 8])
         reshaped = make_tensor_value_info(
             "reshaped", TensorProto.FLOAT, shape_data.int64_data
         )
@@ -151,6 +151,17 @@ class TestStringMethods(unittest.TestCase):
             make_graph([reshape], "reshape", [data, shape], [reshaped], [shape_data])
         )
 
+    def test_concat(self):
+        input1 = make_tensor_value_info("input1", TensorProto.FLOAT, [1, 3, 2, 4])
+        input2 = make_tensor_value_info("input2", TensorProto.FLOAT, [1, 3, 2, 5])
+        output = make_tensor_value_info("output", TensorProto.FLOAT, [1, 3, 2, 9])
+        concat = make_node(
+            "Concat", ["input1", "input2"], ["output"], axis=3, name="concat"
+        )
+        make_and_import_model(
+            make_graph([concat], "concat", [input1, input2], [output])
+        )
+
     # see <https://onnx.ai/onnx/intro/python.html#a-simple-example-a-linear-regression>
     def test_linear(self):
         x = make_tensor_value_info("x", TensorProto.FLOAT, [1, 2, 3])

src/core/graph_handler.cc

@@ -1,5 +1,6 @@
 #include "core/graph_handler.h"
 #include "operators/batch_norm.h"
+#include "operators/concat.h"
 #include "operators/element_wise.h"
 #include "operators/matmul.h"
 #include "operators/reshape.h"
@@ -93,6 +94,15 @@ Tensor GraphHandlerObj::reshape(Tensor data, Tensor reshaped, Shape shape) {
     }
 }
 
+Tensor GraphHandlerObj::concat(TensorVec inputs, Tensor output, int dim) {
+    if (output) {
+        g->addOpWithOutputs<ConcatObj>(std::move(inputs), output, dim);
+        return output;
+    } else {
+        return g->addOp<ConcatObj>(std::move(inputs), output, dim)->getOutput();
+    }
+}
+
 static DataType dtype_repr_convert(int dtype) {
     switch ((OnnxDType)dtype) {
     case OnnxDType::FLOAT:

src/ffi/ffi_infinitensor.cc

@@ -76,6 +76,9 @@ void init_graph_builder(py::module &m) {
              policy::move)
         .def("reshape",
              py::overload_cast<Tensor, Tensor, Shape>(&Handler::reshape),
+             policy::move)
+        .def("concat",
+             py::overload_cast<TensorVec, Tensor, int>(&Handler::concat),
              policy::move);
 }