Translate onnxruntime.md and tensorrt.md (#2320)

Author: blue-q

docs/zh_cn/05-supported-backends/onnxruntime.md

@@ -1,42 +1,42 @@
 # onnxruntime 支持情况
 
-## Introduction of ONNX Runtime
+## ONNX Runtime 介绍
 
-**ONNX Runtime** is a cross-platform inference and training accelerator compatible with many popular ML/DNN frameworks. Check its [github](https://github.com/microsoft/onnxruntime) for more information.
+**ONNX Runtime** 是一个跨平台的推理和训练加速器，与许多流行的ML/DNN框架兼容。查看其[github](https://github.com/microsoft/onnxruntime)以获取更多信息。
 
-## Installation
+## 安装
 
-*Please note that only **onnxruntime>=1.8.1** of on Linux platform is supported by now.*
+*请注意，目前Linux平台只支持 **onnxruntime>=1.8.1** 。*
 
-### Install ONNX Runtime python package
+### 安装ONNX Runtime python包
 
-- CPU Version
+- CPU 版本
 
 ```bash
-pip install onnxruntime==1.8.1 # if you want to use cpu version
+pip install onnxruntime==1.8.1 # 如果你想用cpu版本
 ```
 
-- GPU Version
+- GPU 版本
 
 ```bash
-pip install onnxruntime-gpu==1.8.1 # if you want to use gpu version
+pip install onnxruntime-gpu==1.8.1 # 如果你想用gpu版本
 ```
 
-### Install float16 conversion tool (optional)
+### 安装float16转换工具(可选)
 
-If you want to use float16 precision, install the tool by running the following script:
+如果你想用float16精度，请执行以下脚本安装工具:
 
 ```bash
 pip install onnx onnxconverter-common
 ```
 
-## Build custom ops
+## 构建自定义算子
 
-### Download ONNXRuntime Library
+### 下载ONNXRuntime库
 
-Download `onnxruntime-linux-*.tgz` library from ONNX Runtime [releases](https://github.com/microsoft/onnxruntime/releases/tag/v1.8.1), extract it, expose `ONNXRUNTIME_DIR` and finally add the lib path to `LD_LIBRARY_PATH` as below:
+从ONNX Runtime[发布版本](https://github.com/microsoft/onnxruntime/releases/tag/v1.8.1)下载`onnxruntime-linux-*.tgz`库，并解压，将onnxruntime所在路径添加到`ONNXRUNTIME_DIR`环境变量，最后将lib路径添加到`LD_LIBRARY_PATH`环境变量中，操作如下：
 
-- CPU Version
+- CPU 版本
 
 ```bash
 wget https://github.com/microsoft/onnxruntime/releases/download/v1.8.1/onnxruntime-linux-x64-1.8.1.tgz
@@ -47,7 +47,7 @@ export ONNXRUNTIME_DIR=$(pwd)
 export LD_LIBRARY_PATH=$ONNXRUNTIME_DIR/lib:$LD_LIBRARY_PATH
 ```
 
-- GPU Version
+- GPU 版本
 
 ```bash
 wget https://github.com/microsoft/onnxruntime/releases/download/v1.8.1/onnxruntime-linux-x64-gpu-1.8.1.tgz
@@ -58,49 +58,49 @@ export ONNXRUNTIME_DIR=$(pwd)
 export LD_LIBRARY_PATH=$ONNXRUNTIME_DIR/lib:$LD_LIBRARY_PATH
 ```
 
-### Build on Linux
+### 在Linux上构建
 
-- CPU Version
+- CPU 版本
 
 ```bash
-cd ${MMDEPLOY_DIR} # To MMDeploy root directory
+cd ${MMDEPLOY_DIR} # 进入MMDeploy根目录
 mkdir -p build && cd build
 cmake -DMMDEPLOY_TARGET_DEVICES='cpu' -DMMDEPLOY_TARGET_BACKENDS=ort -DONNXRUNTIME_DIR=${ONNXRUNTIME_DIR} ..
 make -j$(nproc) && make install
 ```
 
-- GPU Version
+- GPU 版本
 
 ```bash
-cd ${MMDEPLOY_DIR} # To MMDeploy root directory
+cd ${MMDEPLOY_DIR} # 进入MMDeploy根目录
 mkdir -p build && cd build
 cmake -DMMDEPLOY_TARGET_DEVICES='cuda' -DMMDEPLOY_TARGET_BACKENDS=ort -DONNXRUNTIME_DIR=${ONNXRUNTIME_DIR} ..
 make -j$(nproc) && make install
 ```
 
-## How to convert a model
+## 如何转换模型
 
-- You could follow the instructions of tutorial [How to convert model](../02-how-to-run/convert_model.md)
+- 你可以按照教程[如何转换模型](../02-how-to-run/convert_model.md)的说明去做
 
-## How to add a new custom op
+## 如何添加新的自定义算子
 
-## Reminder
+## 提示
 
-- The custom operator is not included in [supported operator list](https://github.com/microsoft/onnxruntime/blob/master/docs/OperatorKernels.md) in ONNX Runtime.
-- The custom operator should be able to be exported to ONNX.
+- 自定义算子不包含在ONNX Runtime[支持的算子列表](https://github.com/microsoft/onnxruntime/blob/master/docs/OperatorKernels.md)中。
+- 自定义算子应该能够导出到ONNX。
 
-#### Main procedures
+#### 主要过程
 
-Take custom operator `roi_align` for example.
+以自定义操作符`roi_align`为例。
 
-1. Create a `roi_align` directory in ONNX Runtime source directory `${MMDEPLOY_DIR}/csrc/backend_ops/onnxruntime/`
-2. Add header and source file into `roi_align` directory `${MMDEPLOY_DIR}/csrc/backend_ops/onnxruntime/roi_align/`
-3. Add unit test into `tests/test_ops/test_ops.py`
-   Check [here](../../../tests/test_ops/test_ops.py) for examples.
+1. 在ONNX Runtime源目录`${MMDEPLOY_DIR}/csrc/backend_ops/onnxruntime/`中创建一个`roi_align`目录
+2. 添加头文件和源文件到`roi_align`目录`${MMDEPLOY_DIR}/csrc/backend_ops/onnxruntime/roi_align/`
+3. 将单元测试添加到`tests/test_ops/test_ops.py`中。
+   查看[这里](../../../tests/test_ops/test_ops.py)的例子。
 
-**Finally, welcome to send us PR of adding custom operators for ONNX Runtime in MMDeploy.** :nerd_face:
+\**最后，欢迎发送为MMDeploy添加ONNX Runtime自定义算子的PR。* \*: nerd_face:
 
-## References
+## 参考
 
-- [How to export Pytorch model with custom op to ONNX and run it in ONNX Runtime](https://github.com/onnx/tutorials/blob/master/PyTorchCustomOperator/README.md)
-- [How to add a custom operator/kernel in ONNX Runtime](https://onnxruntime.ai/docs/reference/operators/add-custom-op.html)
+- [如何将具有自定义op的Pytorch模型导出为ONNX并在ONNX Runtime运行](https://github.com/onnx/tutorials/blob/master/PyTorchCustomOperator/README.md)
+- [如何在ONNX Runtime添加自定义算子/内核](https://onnxruntime.ai/docs/reference/operators/add-custom-op.html)

docs/zh_cn/05-supported-backends/tensorrt.md

@@ -1,57 +1,57 @@
 # TensorRT 支持情况
 
-## Installation
+## 安装
 
-### Install TensorRT
+### 安装TensorRT
 
-Please install TensorRT 8 follow [install-guide](https://docs.nvidia.com/deeplearning/tensorrt/install-guide/index.html#installing).
+请按照[安装指南](https://docs.nvidia.com/deeplearning/tensorrt/install-guide/index.html#installing)安装TensorRT8。
 
-**Note**:
+**注意**:
 
-- `pip Wheel File Installation` is not supported yet in this repo.
+- 此版本不支持`pip Wheel File Installation`。
 
-- We strongly suggest you install TensorRT through [tar file](https://docs.nvidia.com/deeplearning/tensorrt/install-guide/index.html#installing-tar)
+- 我们强烈建议通过[tar包](https://docs.nvidia.com/deeplearning/tensorrt/install-guide/index.html#installing-tar)的方式安装TensorRT。
 
-- After installation, you'd better add TensorRT environment variables to bashrc by:
+- 安装完成后，最好通过以下方式将TensorRT环境变量添加到bashrc:
 
   ```bash
-  cd ${TENSORRT_DIR} # To TensorRT root directory
+  cd ${TENSORRT_DIR} # 进入TensorRT根目录
   echo '# set env for TensorRT' >> ~/.bashrc
   echo "export TENSORRT_DIR=${TENSORRT_DIR}" >> ~/.bashrc
   echo 'export LD_LIBRARY_PATH=$TENSORRT_DIR/lib:$TENSORRT_DIR' >> ~/.bashrc
   source ~/.bashrc
   ```
 
-### Build custom ops
+### 构建自定义算子
 
-Some custom ops are created to support models in OpenMMLab, and the custom ops can be built as follow:
+OpenMMLab中创建了一些自定义算子来支持模型，自定义算子可以如下构建:
 
 ```bash
-cd ${MMDEPLOY_DIR} # To MMDeploy root directory
+cd ${MMDEPLOY_DIR} # 进入TensorRT根目录
 mkdir -p build && cd build
 cmake -DMMDEPLOY_TARGET_BACKENDS=trt ..
 make -j$(nproc)
 ```
 
-If you haven't installed TensorRT in the default path, Please add `-DTENSORRT_DIR` flag in CMake.
+如果你没有在默认路径下安装TensorRT，请在CMake中添加`-DTENSORRT_DIR`标志。
 
 ```bash
  cmake -DMMDEPLOY_TARGET_BACKENDS=trt -DTENSORRT_DIR=${TENSORRT_DIR} ..
  make -j$(nproc) && make install
 ```
 
-## Convert model
+## 转换模型
 
-Please follow the tutorial in [How to convert model](../02-how-to-run/convert_model.md). **Note** that the device must be `cuda` device.
+请遵循[如何转换模型](../02-how-to-run/convert_model.md)中的教程。**注意**设备必须是`cuda` 设备。
 
-### Int8 Support
+### Int8 支持
 
-Since TensorRT supports INT8 mode, a custom dataset config can be given to calibrate the model. Following is an example for MMDetection:
+由于TensorRT支持INT8模式，因此可以提供自定义数据集配置来校准模型。MMDetection的示例如下:
 
 ```python
 # calibration_dataset.py
 
-# dataset settings, same format as the codebase in OpenMMLab
+# 数据集设置，格式与OpenMMLab中的代码库相同
 dataset_type = 'CalibrationDataset'
 data_root = 'calibration/dataset/root'
 img_norm_cfg = dict(
@@ -85,32 +85,32 @@ data = dict(
 evaluation = dict(interval=1, metric='bbox')
 ```
 
-Convert your model with this calibration dataset:
+使用此校准数据集转换您的模型:
 
 ```python
 python tools/deploy.py \
     ...
     --calib-dataset-cfg calibration_dataset.py
 ```
 
-If the calibration dataset is not given, the data will be calibrated with the dataset in model config.
+如果没有提供校准数据集，则使用模型配置中的数据集进行校准。
 
 ## FAQs
 
-- Error `Cannot found TensorRT headers` or `Cannot found TensorRT libs`
+- 错误 `Cannot found TensorRT headers`或`Cannot found TensorRT libs`
 
-  Try cmake with flag `-DTENSORRT_DIR`:
+  可以尝试在cmake时使用`-DTENSORRT_DIR`标志:
 
   ```bash
   cmake -DBUILD_TENSORRT_OPS=ON -DTENSORRT_DIR=${TENSORRT_DIR} ..
   make -j$(nproc)
   ```
 
-  Please make sure there are libs and headers in `${TENSORRT_DIR}`.
+  请确保 `${TENSORRT_DIR}`中有库和头文件。
 
-- Error `error: parameter check failed at: engine.cpp::setBindingDimensions::1046, condition: profileMinDims.d[i] <= dimensions.d[i]`
+- 错误 `error: parameter check failed at: engine.cpp::setBindingDimensions::1046, condition: profileMinDims.d[i] <= dimensions.d[i]`
 
-  There is an input shape limit in deployment config:
+  在部署配置中有一个输入形状的限制:
 
   ```python
   backend_config = dict(
@@ -126,14 +126,14 @@ If the calibration dataset is not given, the data will be calibrated with the da
       # other configs
   ```
 
-  The shape of the tensor `input` must be limited between `input_shapes["input"]["min_shape"]` and `input_shapes["input"]["max_shape"]`.
+  `input` 张量的形状必须限制在`input_shapes["input"]["min_shape"]`和`input_shapes["input"]["max_shape"]`之间。
 
-- Error `error: [TensorRT] INTERNAL ERROR: Assertion failed: cublasStatus == CUBLAS_STATUS_SUCCESS`
+- 错误 `error: [TensorRT] INTERNAL ERROR: Assertion failed: cublasStatus == CUBLAS_STATUS_SUCCESS`
 
-  TRT 7.2.1 switches to use cuBLASLt (previously it was cuBLAS). cuBLASLt is the default choice for SM version >= 7.0. However, you may need CUDA-10.2 Patch 1 (Released Aug 26, 2020) to resolve some cuBLASLt issues. Another option is to use the new TacticSource API and disable cuBLASLt tactics if you don't want to upgrade.
+  TRT 7.2.1切换到使用cuBLASLt(以前是cuBLAS)。cuBLASLt是SM版本>= 7.0的默认选择。但是，您可能需要CUDA-10.2补丁1(2020年8月26日发布)来解决一些cuBLASLt问题。如果不想升级，另一个选择是使用新的TacticSource API并禁用cuBLASLt策略。
 
-  Read [this](https://forums.developer.nvidia.com/t/matrixmultiply-failed-on-tensorrt-7-2-1/158187/4) for detail.
+  请阅读[本文](https://forums.developer.nvidia.com/t/matrixmultiply-failed-on-tensorrt-7-2-1/158187/4)了解详情。
 
-- Install mmdeploy on Jetson
+- 在Jetson上安装mmdeploy
 
-  We provide a tutorial to get start on Jetsons [here](../01-how-to-build/jetsons.md).
+  我们在[这里](../01-how-to-build/jetsons.md)提供了一个Jetsons入门教程。