Patch rknpu

source/tnn/device/rknpu/convert/math/rknpu_abs_layer.cc

@@ -0,0 +1,34 @@
+// Tencent is pleased to support the open source community by making TNN available.
+//
+// Copyright (C) 2020 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 "rknpu_unary_operator.h"
+#include "tnn/device/rknpu/convert/rknpu_base_layer.h"
+#include "tnn/device/rknpu/convert/rknpu_utils.h"
+
+namespace TNN_NS {
+
+class RknpuAbsLayer : public RknpuUnaryLayer {
+public:
+    RknpuAbsLayer(LayerType ignore) : RknpuUnaryLayer(LAYER_ABS) {}
+    ~RknpuAbsLayer() {}
+
+protected:
+    Status Convert() {
+        return RknpuUnaryLayer::UnaryConvert(rk::nn::OperatorType::ABS);
+    }
+};
+
+REGISTER_RKNPU_LAYER(Abs, LAYER_ABS);
+
+}  // namespace TNN_NS

source/tnn/device/rknpu/convert/math/rknpu_sub_layer.cc

@@ -0,0 +1,34 @@
+// Tencent is pleased to support the open source community by making TNN available.
+//
+// Copyright (C) 2020 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 "rknpu_binary_layer.h"
+#include "tnn/device/rknpu/convert/rknpu_base_layer.h"
+#include "tnn/device/rknpu/convert/rknpu_utils.h"
+
+namespace TNN_NS {
+
+class RknpuSubLayer : public RknpuBinaryLayer {
+public:
+    RknpuSubLayer(LayerType ignore) : RknpuBinaryLayer(LAYER_SUB) {}
+    ~RknpuSubLayer() {}
+
+protected:
+    Status Convert() {
+        return RknpuBinaryLayer::BinaryConvert(rk::nn::OperatorType::SUBTRACT);
+    }
+};
+
+REGISTER_RKNPU_LAYER(Sub, LAYER_SUB);
+
+}  // namespace TNN_NS

source/tnn/device/rknpu/convert/rknpu_normalize_layer.cc

@@ -0,0 +1,47 @@
+// Tencent is pleased to support the open source community by making TNN available.
+//
+// Copyright (C) 2020 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 <tnn/utils/data_type_utils.h>
+
+#include "rknpu_base_layer.h"
+#include "rknpu_utils.h"
+
+namespace TNN_NS {
+
+DECLARE_RKNPU_LAYER(Normalize, LAYER_NORMALIZE)
+
+Status RknpuNormalizeLayer::Convert() {
+    auto param = dynamic_cast<NormalizeLayerParam *>(param_);
+
+    if (param->axis != 1 || param->across_spatial != 0 || param->p != 2) {
+        return Status(TNNERR_PARAM_ERR, "Error: NormalizeLayer dont support these param!");
+    }
+
+    Status ret = TNN_OK;
+    std::vector<std::shared_ptr<rk::nn::Tensor>> inputs;
+
+    // input
+    inputs.push_back(input_ops_[0]);
+
+    // output
+    ADD_OUTPUT_OP();
+
+    graph_->AddOperator(rk::nn::OperatorType::L2_NORMALIZE, inputs, output_ops_, NULL);
+
+    return ret;
+}
+
+REGISTER_RKNPU_LAYER(Normalize, LAYER_NORMALIZE)
+
+}  // namespace TNN_NS

source/tnn/device/rknpu/convert/rknpu_reduce_mean_layer.cc

@@ -58,6 +58,7 @@ Status RknpuReduceMeanLayer::Convert() {
     for (const auto val : axes) {
         attr.axis.push_back(static_cast<uint32_t>(val));
     }
+    attr.axis_num = attr.axis.size();
     attr.keep_dim = param->keep_dims;
     graph_->AddOperator(rk::nn::OperatorType::REDUCE, inputs, output_ops_, (void *)&attr);
 

source/tnn/device/rknpu/rknpu_network.cc

@@ -314,6 +314,11 @@ Status RknpuNetwork::DeInit() {
     if (blob_manager_)
         delete blob_manager_;
 
+    if (context_ != nullptr) {
+        delete context_;
+        context_ = nullptr;
+    }
+
     return TNN_OK;
 }