# mscnn

* [Source Code](https://github.com/zhaoweicai/mscnn)
* 别人写的[教程](https://gist.github.com/arundasan91/b432cb011d1c45b65222d0fac5f9232c)

## Install

* 从github下载mscnn工程
* 如果git clone出现

  ```
    正克隆到 'mscnn'...
    remote: Counting objects: 1120, done.
    error: RPC failed; curl 56 GnuTLS recv error (-110): The TLS connection was non-properly terminated.
    fatal: The remote end hung up unexpectedly
    fatal: 过早的文件结束符（EOF）
    fatal: index-pack failed
  ```
* 用这种方法避开：

  ```
    mkdir mscnn
    cd mscnn
    git init
    git remote add origin https://github.com/zhaoweicai/mscnn.git
    git pull
  ```
* 我只在ubuntu上安装，所以只讲ubuntu上的依赖:
  * [CUDA和CUDNN](https://www.youtube.com/watch?v=cVWVRA8XXxs)
  * 其他依赖

    ```
    sudo apt-get install libprotobuf-dev libleveldb-dev libsnappy-dev libopencv-dev libhdf5-serial-dev protobuf-compiler
    sudo apt-get install --no-install-recommends libboost-all-dev
    sudo apt-get install libatlas-base-dev
    sudo apt-get install libgflags-dev libgoogle-glog-dev liblmdb-dev
    ```
  * Caffe在debian系列系统上对hdf5的支持导致的[一个坑>>](https://github.com/BVLC/caffe/issues/2347#issuecomment-118508564)

## 编译

```
make all
make pycaffe
make matcaffe
make test
make runtest
```

## 测试（Python版）

* [源代码](https://github.com/GBJim/mscnn/blob/master/examples/caltech/run_mscnn_detection.py)
* 前面是用apt安装的opencv和protobuf，没有安装Python包，所以直接运行会报cv2和protobuf找不到
* 安装opencv python包：

```
sudo apt-get install python-opencv
```

* 这样会安装到系统默认的python解释器中，即`/usr/lib/python2.7/dist-packages`，我们把这里面跟opencv有关的复制到前面编译caffe指定的python解释器目录中，比如`~/anaconda2/lib/python2.7/site-packages`，同时注意改cv2的名字

```
sudo ln -s cv2.x86_64-linux-gnu.so cv2.so
```

搞定opencv的依赖

* 安装protobuf python包: `pip install protobuf`（注意pip要和caffe对应的python解释器绑定）
* 由于前面的那份python代码用了nms来做GPU调用，这个东西是来自py-faster-rcnn的，也是caffe的一个变种，复制[这个目录](https://github.com/rbgirshick/py-faster-rcnn/tree/master/lib/)，然后make，按上面的复制opencv的方法把nms目录复制到caffe对应的python解释器就好了


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://blog.cweihang.io/ml/kit/mscnn.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.