cmdbug / yolov5_ncnn Goto Github PK

🍅 Deploy ncnn on mobile phones. Support Android and iOS. 移动端ncnn部署，支持Android与iOS。

License: GNU General Public License v3.0

Java 0.74% CMake 0.91% C++ 86.97% C 7.92% Objective-C 0.51% Objective-C++ 2.94%

yolo yolov5 ios android ncnn detection object-detection yolov4 yolov3 mobilenet

yolov5_ncnn's Introduction

𝗛𝗲𝗹𝗹𝗼！

┌───┐   ┌───┬───┬───┬───┐ ┌───┬───┬───┬───┐ ┌───┬───┬───┬───┐ ┌───┬───┬───┐
│Esc│   │ F1│ F2│ F3│ F4│ │ F5│ F6│ F7│ F8│ │ F9│F10│F11│F12│ │P/S│S L│P/B│  ┌┐    ┌┐    ┌┐            │
└───┘   └───┴───┴───┴───┘ └───┴───┴───┴───┘ └───┴───┴───┴───┘ └───┴───┴───┘  └┘    └┘    └┘            │
┌───┬───┬───┬───┬───┬───┬───┬───┬───┬───┬───┬───┬───┬───────┐ ┌───┬───┬───┐ ┌───┬───┬───┬───┐          │
│~ `│! 1│@ 2│# 3│$ 4│% 5│^ 6│& 7│* 8│( 9│) 0│_ -│+ =│ BacSp │ │Ins│Hom│PUp│ │N L│ / │ * │ - │   ┌──────┼──────┐
├───┴─┬─┴─┬─┴─┬─┴─┬─┴─┬─┴─┬─┴─┬─┴─┬─┴─┬─┴─┬─┴─┬─┴─┬─┴─┬─────┤ ├───┼───┼───┤ ├───┼───┼───┼───┤   │     ┌┴┐     │
│ Tab │ Q │ W │ E │ R │ T │ Y │ U │ I │ O │ P │{ [│} ]│ | \ │ │Del│End│PDn│ │ 7 │ 8 │ 9 │   │   │     └┬┘     │
├─────┴┬──┴┬──┴┬──┴┬──┴┬──┴┬──┴┬──┴┬──┴┬──┴┬──┴┬──┴┬──┴─────┤ └───┴───┴───┘ ├───┼───┼───┤ + │   ├──────┴──────┤
│ Caps │ A │ S │ D │ F │ G │ H │ J │ K │ L │: ;│" '│ Enter  │               │ 4 │ 5 │ 6 │   │   │             │
├──────┴─┬─┴─┬─┴─┬─┴─┬─┴─┬─┴─┬─┴─┬─┴─┬─┴─┬─┴─┬─┴─┬─┴────────┤     ┌───┐     ├───┼───┼───┼───┤   │             │
│ Shift  │ Z │ X │ C │ V │ B │ N │ M │< ,│> .│? /│  Shift   │     │ ↑ │     │ 1 │ 2 │ 3 │   │   │             │
├─────┬──┴─┬─┴──┬┴───┴───┴───┴───┴───┴──┬┴───┼───┴┬────┬────┤ ┌───┼───┼───┐ ├───┴───┼───┤ E││   │             │
│ Ctrl│    │Alt │         teng          │ Alt│ Fn │    │Ctrl│ │ ← │ ↓ │ → │ │   0   │ . │←─┘│   │     teng    │
└─────┴────┴────┴───────────────────────┴────┴────┴────┴────┘ └───┴───┴───┘ └───────┴───┴───┘   └─────────────┘

yolov5_ncnn's People

Contributors

Stargazers

Watchers

Forkers

tj1116 susyimes cqray1990 zumbalamambo qaz734913414 perfyperfect deeplearning2012 haowangyin hl0071 yangshuanghu qiqihuang fengxingxiang antonizhubar chenyang918 noticeable oaup cuiyong127 sabo-gu dog-qiuqiu zineos anguoyang junan007 20070951 dun933 sinoaidi zhaoyk1986 shengyuting hyaihjq caozhengquan felixzhang7 bytemeow evoluange w1ndseeker qiuweibin2005 wuxiaolianggit wavelet2008 single430 cvding sporterman zhongzebin tcll321 matt45m scangit aliushn yippeesoft cv-ip donaldlee2008 lidaweinuc zzzz737 aniu-hub lucalucashhh llf10811020205 yingunjun fightseed capstone-lol biubiu0719 ixlm p1b234 creama bonima123 duyuankai1992 alanjonson arthurtibame zlszhonglongshen yugongcharley shining-love golfbeta pp-zhangyi xingtaixiaokai johnyfeng daizzhisheng qinhuaping boosting longtengj wenxiang-li rainlolthx dansonc dhenisdj cycloneboys ruofei7 jianyushu szad670401 changerzz kuibu loulansuiye liang-001 aweffr scuizhibin badaniao jeenhyung yuhan19971211 sail-lab155 nabyssache hedilong chenggong0602 adas-eye 93yh youkpan uptodiff apeizou

yolov5_ncnn's Issues

custom yolov4-tiny can inference but no output on yolov5-ncnn-android

大神好，我想用自己訓練的yolov4-tiny 在 android-ncnn 檢測目標
以下是我的yolov4-tiny-opt.param

這是我的模型訓練過程
我用WongKinYiu的yolov4-csp訓練了六個類別的yolov4-tiny
以下是程式碼的網址
https://github.com/WongKinYiu/ScaledYOLOv4/tree/yolov4-csp
參考這個issue
WongKinYiu/ScaledYOLOv4#41
以及更改cfg檔(classes=6 ＆ yolo層輸出的filter=33)
成功訓練出pytorch版本的yolov4-tiny (6-category）
並用ncnn中的darknet2ncnn與ncnnoptimize轉換成ncnn的模型與優化
使用您的android_YOLOV5_NCNN實現推論
但在執行下面這行
ex.extract("output", blob);
一直都沒有輸出結果，想請教大神我可能是哪裡出了問題

新人求教

请问直接下载本代码进入安卓手机，解压安装就可以？

你好，在build的时候报以下错误。。。

Build command failed.
Error while executing process C:\Users\Administrator\AppData\Local\Android\Sdk\cmake\3.10.2.4988404\bin\ninja.exe with arguments {-C C:\Users\Administrator\Desktop\ncnn\YOLOv5_NCNN-master\android_YOLOV5_NCNN\app.cxx\cmake\debug\x86 yolov5}
ninja: Entering directory `C:\Users\Administrator\Desktop\ncnn\YOLOv5_NCNN-master\android_YOLOV5_NCNN\app.cxx\cmake\debug\x86'

ninja: error: 'C:/Users/Administrator/Desktop/ncnn/YOLOv5_NCNN-master/android_YOLOV5_NCNN/app/src/main/cpp/opencv/x86/libopencv_java4.so', needed by 'C:/Users/Administrator/Desktop/ncnn/YOLOv5_NCNN-master/android_YOLOV5_NCNN/app/build/intermediates/cmake/debug/obj/x86/libyolov5.so', missing and no known rule to make it

请问一下这个问题，是需要在cmake里指定吗，我改成了ncnn或者ncnnvulkan都不行

C/C++ debug|armeabi-v7a : CMake Error at app\src\main\cpp\CMakeLists.txt:103 (target_link_libraries):
Cannot specify link libraries for target "yolov5" which is not built by
this project.

用的android4.0那个版本啊？Could not resolve androidx.constraintlayout:constraintlayout:1.1.3

Merge our project? YOLOV4 tiny + lane detection on Android with 8 FPS!

YOLOV4 tiny + lane detection on Android with 8 FPS!

Tested with HONOR 20PRO
Test video :

https://www.bilibili.com/video/BV1xv411e7zJ/

https://www.bilibili.com/video/BV18K4y1n7Ht/

Published at :
http://hayoou.com/adas

APP download ：https://pan.baidu.com/s/1Q-huWTcHGWfyD8AXxVoyow
code：adas

open souce :

https://github.com/youkpan/hayoou_safe_driving_android

测试

您好，请问我在测试项目时会提示找不到libopencv_java4.so

can you share the android apk file? thanks

大佬可以单独写个ncnn的yolov5代码么？不是android和ios的，谢谢！！！

自己从头到尾转换了一下yolov5s.pt到yolov5s.param和yolov5s.bin文件，在android studio中替换后没有检测结果，并且几秒钟后会退回到主界面，会是什原因呢？

2021-03-10 17:08:28.596 5661-5661/? I/com.wzt.yolov5: Not late-enabling -Xcheck:jni (already on)
2021-03-10 17:08:28.639 5661-5661/? I/com.wzt.yolov5: Unquickening 12 vdex files!
2021-03-10 17:08:28.643 5661-5661/? W/com.wzt.yolov5: Unexpected CPU variant for X86 using defaults: x86
2021-03-10 17:08:28.648 5661-5661/? I/ndk_translation: Initialized NDK translation (armv7l), version 0.2.2
2021-03-10 17:08:28.921 5661-5661/com.wzt.yolov5 D/ApplicationLoaders: Returning zygote-cached class loader: /system/framework/android.test.base.jar
2021-03-10 17:08:29.066 5661-5661/com.wzt.yolov5 D/NetworkSecurityConfig: No Network Security Config specified, using platform default
2021-03-10 17:08:29.066 5661-5661/com.wzt.yolov5 D/NetworkSecurityConfig: No Network Security Config specified, using platform default
2021-03-10 17:08:29.070 5661-5661/com.wzt.yolov5 I/MultiDex: VM with version 2.1.0 has multidex support
2021-03-10 17:08:29.070 5661-5661/com.wzt.yolov5 I/MultiDex: Installing application
2021-03-10 17:08:29.070 5661-5661/com.wzt.yolov5 I/MultiDex: VM has multidex support, MultiDex support library is disabled.
2021-03-10 17:08:29.071 5661-5661/com.wzt.yolov5 D/Camera2Initializer: CameraX initializing with Camera2 ...
2021-03-10 17:08:29.078 5661-5661/com.wzt.yolov5 I/CameraManagerGlobal: Connecting to camera service
2021-03-10 17:08:29.096 5661-5661/com.wzt.yolov5 D/CameraRepository: Added camera: 0
2021-03-10 17:08:29.109 5661-5697/com.wzt.yolov5 D/UseCaseAttachState: Active and online use case: [] for camera: 0
2021-03-10 17:08:29.110 5661-5661/com.wzt.yolov5 D/CameraRepository: Added camera: 1
2021-03-10 17:08:29.113 5661-5697/com.wzt.yolov5 D/UseCaseAttachState: Active and online use case: [] for camera: 1
2021-03-10 17:08:29.127 5661-5700/com.wzt.yolov5 D/libEGL: loaded /vendor/lib/egl/libEGL_emulation.so
2021-03-10 17:08:29.130 5661-5700/com.wzt.yolov5 D/libEGL: loaded /vendor/lib/egl/libGLESv1_CM_emulation.so
2021-03-10 17:08:29.133 5661-5700/com.wzt.yolov5 D/libEGL: loaded /vendor/lib/egl/libGLESv2_emulation.so
2021-03-10 17:08:29.202 5661-5661/com.wzt.yolov5 W/com.wzt.yolov5: Accessing hidden method Landroid/view/View;->computeFitSystemWindows(Landroid/graphics/Rect;Landroid/graphics/Rect;)Z (greylist, reflection, allowed)
2021-03-10 17:08:29.202 5661-5661/com.wzt.yolov5 W/com.wzt.yolov5: Accessing hidden method Landroid/view/ViewGroup;->makeOptionalFitsSystemWindows()V (greylist, reflection, allowed)
2021-03-10 17:08:29.365 5661-5698/com.wzt.yolov5 D/HostConnection: HostConnection::get() New Host Connection established 0xf1a179a0, tid 5698
2021-03-10 17:08:29.370 5661-5698/com.wzt.yolov5 D/HostConnection: HostComposition ext ANDROID_EMU_CHECKSUM_HELPER_v1 ANDROID_EMU_native_sync_v2 ANDROID_EMU_native_sync_v3 ANDROID_EMU_native_sync_v4 ANDROID_EMU_dma_v1 ANDROID_EMU_direct_mem ANDROID_EMU_host_composition_v1 ANDROID_EMU_host_composition_v2 ANDROID_EMU_vulkan ANDROID_EMU_deferred_vulkan_commands ANDROID_EMU_vulkan_null_optional_strings ANDROID_EMU_vulkan_create_resources_with_requirements ANDROID_EMU_YUV_Cache ANDROID_EMU_async_unmap_buffer ANDROID_EMU_vulkan_ignored_handles ANDROID_EMU_vulkan_free_memory_sync ANDROID_EMU_vulkan_shader_float16_int8 ANDROID_EMU_vulkan_async_queue_submit GL_OES_EGL_image_external_essl3 GL_OES_vertex_array_object GL_KHR_texture_compression_astc_ldr ANDROID_EMU_host_side_tracing ANDROID_EMU_async_frame_commands ANDROID_EMU_gles_max_version_3_0
2021-03-10 17:08:29.372 5661-5698/com.wzt.yolov5 W/OpenGLRenderer: Failed to choose config with EGL_SWAP_BEHAVIOR_PRESERVED, retrying without...
2021-03-10 17:08:29.386 5661-5698/com.wzt.yolov5 D/EGL_emulation: eglCreateContext: 0xf1a18260: maj 3 min 0 rcv 3
2021-03-10 17:08:29.404 5661-5698/com.wzt.yolov5 D/EGL_emulation: eglMakeCurrent: 0xf1a18260: ver 3 0 (tinfo 0xf1d32430) (first time)
2021-03-10 17:08:29.419 5661-5698/com.wzt.yolov5 I/Gralloc4: mapper 4.x is not supported
2021-03-10 17:08:29.446 5661-5698/com.wzt.yolov5 D/HostConnection: createUnique: call
2021-03-10 17:08:29.446 5661-5698/com.wzt.yolov5 D/HostConnection: HostConnection::get() New Host Connection established 0xf1a173f0, tid 5698
2021-03-10 17:08:29.446 5661-5698/com.wzt.yolov5 D/goldfish-address-space: allocate: Ask for block of size 0x100
2021-03-10 17:08:29.446 5661-5698/com.wzt.yolov5 D/goldfish-address-space: allocate: ioctl allocate returned offset 0x3fda03000 size 0x2000
2021-03-10 17:08:29.450 5661-5698/com.wzt.yolov5 D/HostConnection: HostComposition ext ANDROID_EMU_CHECKSUM_HELPER_v1 ANDROID_EMU_native_sync_v2 ANDROID_EMU_native_sync_v3 ANDROID_EMU_native_sync_v4 ANDROID_EMU_dma_v1 ANDROID_EMU_direct_mem ANDROID_EMU_host_composition_v1 ANDROID_EMU_host_composition_v2 ANDROID_EMU_vulkan ANDROID_EMU_deferred_vulkan_commands ANDROID_EMU_vulkan_null_optional_strings ANDROID_EMU_vulkan_create_resources_with_requirements ANDROID_EMU_YUV_Cache ANDROID_EMU_async_unmap_buffer ANDROID_EMU_vulkan_ignored_handles ANDROID_EMU_vulkan_free_memory_sync ANDROID_EMU_vulkan_shader_float16_int8 ANDROID_EMU_vulkan_async_queue_submit GL_OES_EGL_image_external_essl3 GL_OES_vertex_array_object GL_KHR_texture_compression_astc_ldr ANDROID_EMU_host_side_tracing ANDROID_EMU_async_frame_commands ANDROID_EMU_gles_max_version_3_0

How make train Ultralight-Nano-SimplePose.bin?

Where you train this model?
Thanks you

iOS demo ncnn/ncnn/net.h和ncnn/net都是not found

有没有相关的原始模型

我最近在看你项目上的DBFace内容，请问有没有dbface.bin和dbface.param对应的pth文件。

请问这个什么问题

CmakeList里的enable rtti and exceptions的作用

请问upCmakeList里的enable rtti and exceptions的作用是什么啊？

Question:Real-time detection method using a camera

Is real-time detection in the camera using your own model possible?

ocr-ncnn报错。。。

android_YOLOV5_NCNN/app/src/main/cpp/ocr/RRLib.cpp:223: error: undefined reference to 'cv::_InputArray::_InputArray(cv::MatExpr const&)'

大佬，玩你部署的ocr-lite的时候，报这个错。。帮忙看下

安卓编译成功，在手机上运行打开图片报错，调用摄像头没有问题

你好，我在运行您的项目时使用摄像头是没有问题的

当我使用 ChineseOCR 选择图片报了以下的错误，应该是图片没有正确 get

这里我去看240行，然后断点调试，出现了下面这个信息，权限问题？？，由于我不熟悉安卓的开发，这里便不知改如何修改了，希望可以得到您@WZTENG 的帮助！！

环境：真机小米8, 安卓10

我训练了yolov5自己的模型，类别是1，在转成ncnn替换了YOLOv5_NCNN-master\iOS_YOLOv5NCNN\YOLOv5NCNN\res\

我训练了yolov5自己的模型，类别是1，在转成ncnn替换了YOLOv5_NCNN-master\iOS_YOLOv5NCNN\YOLOv5NCNN\res\yolov5.bin 和yolov5.param之后执行demo报错
2021-01-18 11:00:43.932899+0800 NCNN Demo[9473:20535076] new YoloV5
layer Shape_246 not exists or registered
我用的是最新版本的yolov5，version3训练的，我想知道作者yolov5demo用的是那个版本的？还有一个就是我修改了训练类别，这个直接替换文件还需要做什么操作吗？

Hardocded SimplePose Mean and Norm values, can you explain them

Hello, I was wondering if anyone can give any information about the hard coded values of MEAN and NORM about Simple Pose model, for both person and pose detector, this values are:

Person Detector

 float norm[3] = {1 / 255.f, 1 / 255.f, 1 / 255.f};
 float mean[3] = {0, 0, 0};

Pose Detector

const float mean_vals[3] = {0.485f * 255.f, 0.456f * 255.f, 0.406f * 255.f};
const float norm_vals[3] = {1 / 0.229f / 255.f, 1 / 0.224f / 255.f, 1 / 0.225f / 255.f};

Thank you!

Why need convert from RGB to BGR

Hello, i noticed that preprocess() in NanoDet.cpp, there is a line of code :
in = ncnn::Mat::from_android_bitmap_resize(env, image, ncnn::Mat::PIXEL_RGBA2BGR, input_size, input_size);
I want to know why we need to convert RGB to BGR, ncnn's default channel order is BGR ? Thanks in advance !

Do I need to download any other files to make it work?
Help would be much appreciated.