YOLOv4: Darknet 如何於 Docker 編譯，及訓練 COCO 子集

GoCodingInMyWay發表於2020-09-11

YOLODocker編譯

YOLO 演算法是非常著名的目標檢測演算法。從其全稱 You Only Look Once: Unified, Real-Time Object Detection ，可以看出它的特性：

Look Once: one-stage (one-shot object detectors) 演算法，把目標檢測的兩個任務分類和定位一步完成。
Unified: 統一的架構，提供 end-to-end 的訓練和預測。
Real-Time: 實時性，初代論文給出的指標 FPS 45 , mAP 63.4 。

YOLOv4: Optimal Speed and Accuracy of Object Detection ，於今年 4 月公佈，採用了很多近些年 CNN 領域優秀的優化技巧。其平衡了精度與速度，目前在實時目標檢測演算法中精度是最高的。

論文地址：

YOLO: https://arxiv.org/abs/1506.02640
YOLO v4: https://arxiv.org/abs/2004.10934

原始碼地址：

YOLO: https://github.com/pjreddie/darknet
YOLO v4: https://github.com/AlexeyAB/darknet

本文將介紹 YOLOv4 官方 Darknet 實現，如何於 Docker 編譯使用。以及從 MS COCO 2017 資料集中怎麼選出部分物體，訓練出模型。

主要內容有：

準備 Docker 映象
準備 COCO 資料集
用預訓練模型進行推斷
準備 COCO 資料子集
訓練自己的模型並推斷
參考內容

準備 Docker 映象

首先，準備 Docker ，請見：Docker: Nvidia Driver, Nvidia Docker 推薦安裝步驟。

之後，開始準備映象，從下到上的層級為：

nvidia/cuda: https://hub.docker.com/r/nvidia/cuda
OpenCV: https://github.com/opencv/opencv
Darknet: https://github.com/AlexeyAB/darknet

nvidia/cuda

準備 Nvidia 基礎 CUDA 映象。這裡我們選擇 CUDA 10.2 ，不用最新 CUDA 11，因為現在 PyTorch 等都還都是 10.2 呢。

拉取映象：

docker pull nvidia/cuda:10.2-cudnn7-devel-ubuntu18.04

測試映象：

$ docker run --gpus all nvidia/cuda:10.2-cudnn7-devel-ubuntu18.04 nvidia-smi
Sun Aug  8 00:00:00 2020
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 440.100      Driver Version: 440.100      CUDA Version: 10.2     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  GeForce RTX 208...  Off  | 00000000:07:00.0  On |                  N/A |
|  0%   48C    P8    14W / 300W |    340MiB / 11016MiB |      2%      Default |
+-------------------------------+----------------------+----------------------+
|   1  GeForce RTX 208...  Off  | 00000000:08:00.0 Off |                  N/A |
|  0%   45C    P8    19W / 300W |      1MiB / 11019MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |
|=============================================================================|
+-----------------------------------------------------------------------------+

OpenCV

基於 nvidia/cuda 映象，構建 OpenCV 的映象：

cd docker/ubuntu18.04-cuda10.2/opencv4.4.0/

docker build \
-t joinaero/ubuntu18.04-cuda10.2:opencv4.4.0 \
--build-arg opencv_ver=4.4.0 \
--build-arg opencv_url=https://gitee.com/cubone/opencv.git \
--build-arg opencv_contrib_url=https://gitee.com/cubone/opencv_contrib.git \
.

其 Dockerfile 可見這裡： https://github.com/ikuokuo/start-yolov4/blob/master/docker/ubuntu18.04-cuda10.2/opencv4.4.0/Dockerfile 。

Darknet

基於 OpenCV 映象，構建 Darknet 映象：

cd docker/ubuntu18.04-cuda10.2/opencv4.4.0/darknet/

docker build \
-t joinaero/ubuntu18.04-cuda10.2:opencv4.4.0-darknet \
.

其 Dockerfile 可見這裡： https://github.com/ikuokuo/start-yolov4/blob/master/docker/ubuntu18.04-cuda10.2/opencv4.4.0/darknet/Dockerfile 。

上述映象已上傳 Docker Hub 。如果 Nvidia 驅動能夠支援 CUDA 10.2 ，那可以直接拉取該映象：

docker pull joinaero/ubuntu18.04-cuda10.2:opencv4.4.0-darknet

準備 COCO 資料集

MS COCO 2017 下載地址: http://cocodataset.org/#download

影像，包括：

2017 Train images [118K/18GB]
- http://images.cocodataset.org/zips/train2017.zip
2017 Val images [5K/1GB]
- http://images.cocodataset.org/zips/val2017.zip
2017 Test images [41K/6GB]
- http://images.cocodataset.org/zips/test2017.zip
2017 Unlabeled images [123K/19GB]
- http://images.cocodataset.org/zips/unlabeled2017.zip

標註，包括：

2017 Train/Val annotations [241MB]
- http://images.cocodataset.org/annotations/annotations_trainval2017.zip
2017 Stuff Train/Val annotations [1.1GB]
- http://images.cocodataset.org/annotations/stuff_annotations_trainval2017.zip
2017 Panoptic Train/Val annotations [821MB]
- http://images.cocodataset.org/annotations/panoptic_annotations_trainval2017.zip
2017 Testing Image info [1MB]
- http://images.cocodataset.org/annotations/image_info_test2017.zip
2017 Unlabeled Image info [4MB]
- http://images.cocodataset.org/annotations/image_info_unlabeled2017.zip

用預訓練模型進行推斷

預訓練模型 yolov4.weights ，下載地址 https://github.com/AlexeyAB/darknet/releases/download/darknet_yolo_v3_optimal/yolov4.weights 。

執行映象：

xhost +local:docker

docker run -it --gpus all \
-e DISPLAY \
-e QT_X11_NO_MITSHM=1 \
-v /tmp/.X11-unix:/tmp/.X11-unix \
-v $HOME/.Xauthority:/root/.Xauthority \
--name darknet \
--mount type=bind,source=$HOME/Codes/devel/datasets/coco2017,target=/home/coco2017 \
--mount type=bind,source=$HOME/Codes/devel/models/yolov4,target=/home/yolov4 \
joinaero/ubuntu18.04-cuda10.2:opencv4.4.0-darknet

進行推斷：

./darknet detector test cfg/coco.data cfg/yolov4.cfg /home/yolov4/yolov4.weights \
-thresh 0.25 -ext_output -show -out /home/coco2017/result.json \
/home/coco2017/test2017/000000000001.jpg

推斷結果：

準備 COCO 資料子集

MS COCO 2017 資料集有 80 個物體標籤。我們從中選取自己關注的物體，重組個子資料集。

首先，獲取樣例程式碼：

git clone https://github.com/ikuokuo/start-yolov4.git

scripts/coco2yolo.py: COCO 資料集轉 YOLO 資料集的指令碼
scripts/coco/label.py: COCO 資料集的物體標籤有哪些
cfg/coco/coco.names: 編輯我們想要的那些物體標籤

之後，準備資料集：

cd start-yolov4/
pip install -r scripts/requirements.txt

export COCO_DIR=$HOME/Codes/devel/datasets/coco2017

# train
python scripts/coco2yolo.py \
--coco_img_dir $COCO_DIR/train2017/ \
--coco_ann_file $COCO_DIR/annotations/instances_train2017.json \
--yolo_names_file ./cfg/coco/coco.names \
--output_dir ~/yolov4/coco2017/ \
--output_name train2017 \
--output_img_prefix /home/yolov4/coco2017/train2017/

# valid
python scripts/coco2yolo.py \
--coco_img_dir $COCO_DIR/val2017/ \
--coco_ann_file $COCO_DIR/annotations/instances_val2017.json \
--yolo_names_file ./cfg/coco/coco.names \
--output_dir ~/yolov4/coco2017/ \
--output_name val2017 \
--output_img_prefix /home/yolov4/coco2017/val2017/

資料集，內容如下：

~/yolov4/coco2017/
├── train2017/
│   ├── 000000000071.jpg
│   ├── 000000000071.txt
│   ├── ...
│   ├── 000000581899.jpg
│   └── 000000581899.txt
├── train2017.txt
├── val2017/
│   ├── 000000001353.jpg
│   ├── 000000001353.txt
│   ├── ...
│   ├── 000000579818.jpg
│   └── 000000579818.txt
└── val2017.txt

訓練自己的模型並推斷

準備必要檔案

cfg/coco/coco.names <cfg/coco/coco.names.bak has original 80 objects>
- Edit: keep desired objects
cfg/coco/yolov4.cfg <cfg/coco/yolov4.cfg.bak is original file>
- Download yolov4.cfg, then changed:
- batch=64, subdivisions=32 <32 for 8-12 GB GPU-VRAM>
- width=512, height=512 <any value multiple of 32>
- classes=<your number of objects in each of 3 [yolo]-layers>
- max_batches=<classes*2000, but not less than number of training images and not less than 6000>
- steps=<max_batches*0.8, max_batches*0.9>
- filters=<(classes+5)x3, in the 3 [convolutional] before each [yolo] layer>
- ~~filters=<(classes+9)x3, in the 3 [convolutional] before each [Gaussian_yolo] layer>~~
cfg/coco/coco.data
- Edit: train, valid to YOLO datas

csdarknet53-omega.conv.105

Download csdarknet53-omega_final.weights, then run:

docker run -it --rm --gpus all \
--mount type=bind,source=$HOME/Codes/devel/models/yolov4,target=/home/yolov4 \
joinaero/ubuntu18.04-cuda10.2:opencv4.4.0-darknet

./darknet partial cfg/csdarknet53-omega.cfg /home/yolov4/csdarknet53-omega_final.weights /home/yolov4/csdarknet53-omega.conv.105 105

訓練自己的模型

執行映象：

cd start-yolov4/

xhost +local:docker

docker run -it --gpus all \
-e DISPLAY \
-e QT_X11_NO_MITSHM=1 \
-v /tmp/.X11-unix:/tmp/.X11-unix \
-v $HOME/.Xauthority:/root/.Xauthority \
--name darknet \
--mount type=bind,source=$HOME/Codes/devel/models/yolov4,target=/home/yolov4 \
--mount type=bind,source=$HOME/yolov4/coco2017,target=/home/yolov4/coco2017 \
--mount type=bind,source=$PWD/cfg/coco,target=/home/cfg \
joinaero/ubuntu18.04-cuda10.2:opencv4.4.0-darknet

進行訓練：

mkdir -p /home/yolov4/coco2017/backup

# Training command
./darknet detector train /home/cfg/coco.data /home/cfg/yolov4.cfg /home/yolov4/csdarknet53-omega.conv.105 -map

中途可以中斷訓練，然後這樣繼續：

# Continue training
./darknet detector train /home/cfg/coco.data /home/cfg/yolov4.cfg /home/yolov4/coco2017/backup/yolov4_last.weights -map

yolov4_last.weights 每迭代 100 次，會被記錄。

如果多 GPU 訓練，可以在 1000 次迭代後，加引數 -gpus 0,1 ，再繼續：

# How to train with multi-GPU
# 1. Train it first on 1 GPU for like 1000 iterations
# 2. Then stop and by using partially-trained model `/backup/yolov4_1000.weights` run training with multigpu
./darknet detector train /home/cfg/coco.data /home/cfg/yolov4.cfg /home/yolov4/coco2017/backup/yolov4_1000.weights -gpus 0,1 -map

訓練過程，記錄如下：

加引數 -map 後，上圖會顯示有紅線 mAP。

檢視模型 mAP@IoU=50 精度：

$ ./darknet detector map /home/cfg/coco.data /home/cfg/yolov4.cfg /home/yolov4/coco2017/backup/yolov4_final.weights
...
Loading weights from /home/yolov4/coco2017/backup/yolov4_final.weights...
 seen 64, trained: 384 K-images (6 Kilo-batches_64)
Done! Loaded 162 layers from weights-file

 calculation mAP (mean average precision)...
 Detection layer: 139 - type = 27
 Detection layer: 150 - type = 27
 Detection layer: 161 - type = 27
160
 detections_count = 745, unique_truth_count = 190
class_id = 0, name = train, ap = 80.61%   	 (TP = 142, FP = 18)

 for conf_thresh = 0.25, precision = 0.89, recall = 0.75, F1-score = 0.81
 for conf_thresh = 0.25, TP = 142, FP = 18, FN = 48, average IoU = 75.31 %

 IoU threshold = 50 %, used Area-Under-Curve for each unique Recall
 mean average precision (mAP@0.50) = 0.806070, or 80.61 %
Total Detection Time: 4 Seconds