[RKNPU2]support rknpu2 ClasModel #957 (#964)

* [RKNPU2]support rknpu2 ClasModel #957 * [RKNPU2]support rknpu2 ClasModel #957 * [RKNPU2]support rknpu2 add Resnet50_vd example #957 * [RKNPU2]support rknpu2 add Resnet50_vd example #957 * [RKNPU2]support rknpu2, improve doc #957 * [RKNPU2]support rknpu2, improve doc #957 * [RKNPU2]support rknpu2, improve doc #957 * [RKNPU2]support rknpu2, improve doc #957 * [RKNPU2]support rknpu2, improve doc #957 * [RKNPU2]support rknpu2, improve doc #957 * [RKNPU2]support rknpu2, improve doc #957
2025-11-03 11:02:01 +08:00 · 2022-12-28 17:58:18 +08:00
parent 02425bf280
commit 973c746d06
13 changed files with 390 additions and 17 deletions
--- a/examples/vision/classification/paddleclas/rknpu2/python/README.md
+++ b/examples/vision/classification/paddleclas/rknpu2/python/README.md
@@ -0,0 +1,35 @@
+# PaddleClas Python部署示例
+
+在部署前，需确认以下两个步骤
+
+- 1. 软硬件环境满足要求，参考[FastDeploy环境要求](../../../../../../docs/cn/build_and_install/rknpu2.md)
+
+本目录下提供`infer.py`快速完成 ResNet50_vd 在RKNPU上部署的示例。执行如下脚本即可完成
+
+```bash
+# 下载部署示例代码
+git clone https://github.com/PaddlePaddle/FastDeploy.git
+cd FastDeploy/examples/vision/classification/paddleclas/rknpu2/python
+
+# 下载图片
+wget https://gitee.com/paddlepaddle/PaddleClas/raw/release/2.4/deploy/images/ImageNet/ILSVRC2012_val_00000010.jpeg
+
+# 推理
+python3 infer.py --model_file ./ResNet50_vd_infer/ResNet50_vd_infer_rk3588.rknn  --config_file ResNet50_vd_infer/inference_cls.yaml  --image ILSVRC2012_val_00000010.jpeg
+
+# 运行完成后返回结果如下所示
+ClassifyResult(
+label_ids: 153, 
+scores: 0.684570, 
+)
+```
+
+
+## 注意事项
+RKNPU上对模型的输入要求是使用NHWC格式，且图片归一化操作会在转RKNN模型时，内嵌到模型中，因此我们在使用FastDeploy部署时，
+DisablePermute(C++)或`disable_permute(Python)，在预处理阶段禁用数据格式的转换。
+
+## 其它文档
+- [ResNet50_vd C++部署](../cpp)
+- [模型预测结果说明](../../../../../../docs/api/vision_results/)
+- [转换ResNet50_vd RKNN模型文档](../README.md)
--- a/examples/vision/classification/paddleclas/rknpu2/python/infer.py
+++ b/examples/vision/classification/paddleclas/rknpu2/python/infer.py
@@ -0,0 +1,50 @@
+# Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# 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 fastdeploy as fd
+import cv2
+import os
+
+
+def parse_arguments():
+    import argparse
+    import ast
+    parser = argparse.ArgumentParser()
+    parser.add_argument(
+        "--model_file", required=True, help="Path of rknn model.")
+    parser.add_argument("--config_file", required=True, help="Path of config.")
+    parser.add_argument(
+        "--image", type=str, required=True, help="Path of test image file.")
+    return parser.parse_args()
+
+
+if __name__ == "__main__":
+    args = parse_arguments()
+
+    model_file = args.model_file
+    params_file = ""
+    config_file = args.config_file
+    # 配置runtime，加载模型
+    runtime_option = fd.RuntimeOption()
+    runtime_option.use_rknpu2()
+    model = fd.vision.classification.ResNet50vd(
+        model_file,
+        params_file,
+        config_file,
+        runtime_option=runtime_option,
+        model_format=fd.ModelFormat.RKNN)
+    # 禁用通道转换    
+    model.preprocessor.disable_permute()
+    im = cv2.imread(args.image)
+    result = model.predict(im, topk=1)
+    print(result)