Merged from Alexey/darknet

README.md

@@ -4,7 +4,7 @@
 
 Paper YOLO v4: https://arxiv.org/abs/2004.10934
 
-Paper Scaled YOLO v4: https://arxiv.org/abs/2011.08036  use to reproduce results: [ScaledYOLOv4](https://github.com/WongKinYiu/ScaledYOLOv4)
+Paper Scaled YOLO v4: * **[CVPR 2021](https://openaccess.thecvf.com/content/CVPR2021/html/Wang_Scaled-YOLOv4_Scaling_Cross_Stage_Partial_Network_CVPR_2021_paper.html)**: use to reproduce results: [ScaledYOLOv4](https://github.com/WongKinYiu/ScaledYOLOv4)
 
 More details in articles on medium:
 
@@ -152,6 +152,12 @@ FPS on RTX 2070 (R) and Tesla V100 (V):
 - [yolov4-p5.cfg](https://raw.githubusercontent.com/AlexeyAB/darknet/master/cfg/yolov4-p5.cfg) - 896x896 - **70.0% [email protected] (51.6% [email protected]:0.95) - 43(V) FPS** - xxx BFlops (xxx FMA) - 271 MB: [yolov4-p5.weights](https://github.com/AlexeyAB/darknet/releases/download/darknet_yolo_v4_pre/yolov4-p5.weights)
   - pre-trained weights for training: https://github.com/AlexeyAB/darknet/releases/download/darknet_yolo_v4_pre/yolov4-p5.conv.232
 
+- [yolov4-csp-x-swish.cfg](https://raw.githubusercontent.com/AlexeyAB/darknet/master/cfg/yolov4-csp-x-swish.cfg) - 640x640 - **69.9% [email protected] (51.5% [email protected]:0.95) - 23(R) FPS / 50(V) FPS** - 221 BFlops (110 FMA) - 381 MB: [yolov4-csp-x-swish.weights](https://github.com/AlexeyAB/darknet/releases/download/darknet_yolo_v4_pre/yolov4-csp-x-swish.weights)
+  - pre-trained weights for training: https://github.com/AlexeyAB/darknet/releases/download/darknet_yolo_v4_pre/yolov4-csp-x-swish.conv.192
+
+- [yolov4-csp-swish.cfg](https://raw.githubusercontent.com/AlexeyAB/darknet/master/cfg/yolov4-csp-swish.cfg) - 640x640 - **68.7% [email protected] (50.0% [email protected]:0.95) - 70(V) FPS** - 120 (60 FMA) - 202 MB: [yolov4-csp-swish.weights](https://github.com/AlexeyAB/darknet/releases/download/darknet_yolo_v4_pre/yolov4-csp-swish.weights)
+  - pre-trained weights for training: https://github.com/AlexeyAB/darknet/releases/download/darknet_yolo_v4_pre/yolov4-csp-swish.conv.164
+
 - [yolov4x-mish.cfg](https://raw.githubusercontent.com/AlexeyAB/darknet/master/cfg/yolov4x-mish.cfg) - 640x640 - **68.5% [email protected] (50.1% [email protected]:0.95) - 23(R) FPS / 50(V) FPS** - 221 BFlops (110 FMA) - 381 MB: [yolov4x-mish.weights](https://github.com/AlexeyAB/darknet/releases/download/darknet_yolo_v4_pre/yolov4x-mish.weights)
   - pre-trained weights for training: https://github.com/AlexeyAB/darknet/releases/download/darknet_yolo_v4_pre/yolov4x-mish.conv.166
 

darknet.py

@@ -24,6 +24,7 @@ class BOX(Structure):
 class DETECTION(Structure):
     _fields_ = [("bbox", BOX),
                 ("classes", c_int),
+                ("best_class_idx", c_int),
                 ("prob", POINTER(c_float)),
                 ("mask", POINTER(c_float)),
                 ("objectness", c_float),
@@ -133,6 +134,56 @@ def decode_detection(detections):
         decoded.append((str(label), confidence, bbox))
     return decoded
 
+# https://www.pyimagesearch.com/2015/02/16/faster-non-maximum-suppression-python/
+# Malisiewicz et al.
+def non_max_suppression_fast(detections, overlap_thresh):
+    boxes = []
+    for detection in detections:
+        _, _, _, (x, y, w, h) = detection
+        x1 = x - w / 2
+        y1 = y - h / 2
+        x2 = x + w / 2
+        y2 = y + h / 2
+        boxes.append(np.array([x1, y1, x2, y2]))
+    boxes_array = np.array(boxes)
+
+    # initialize the list of picked indexes
+    pick = []
+    # grab the coordinates of the bounding boxes
+    x1 = boxes_array[:, 0]
+    y1 = boxes_array[:, 1]
+    x2 = boxes_array[:, 2]
+    y2 = boxes_array[:, 3]
+    # compute the area of the bounding boxes and sort the bounding
+    # boxes by the bottom-right y-coordinate of the bounding box
+    area = (x2 - x1 + 1) * (y2 - y1 + 1)
+    idxs = np.argsort(y2)
+    # keep looping while some indexes still remain in the indexes
+    # list
+    while len(idxs) > 0:
+        # grab the last index in the indexes list and add the
+        # index value to the list of picked indexes
+        last = len(idxs) - 1
+        i = idxs[last]
+        pick.append(i)
+        # find the largest (x, y) coordinates for the start of
+        # the bounding box and the smallest (x, y) coordinates
+        # for the end of the bounding box
+        xx1 = np.maximum(x1[i], x1[idxs[:last]])
+        yy1 = np.maximum(y1[i], y1[idxs[:last]])
+        xx2 = np.minimum(x2[i], x2[idxs[:last]])
+        yy2 = np.minimum(y2[i], y2[idxs[:last]])
+        # compute the width and height of the bounding box
+        w = np.maximum(0, xx2 - xx1 + 1)
+        h = np.maximum(0, yy2 - yy1 + 1)
+        # compute the ratio of overlap
+        overlap = (w * h) / area[idxs[:last]]
+        # delete all indexes from the index list that have
+        idxs = np.delete(idxs, np.concatenate(([last],
+                                               np.where(overlap > overlap_thresh)[0])))
+        # return only the bounding boxes that were picked using the
+        # integer data type
+    return [detections[i] for i in pick]
 
 def remove_negatives(detections, class_names, num):
     """
@@ -148,6 +199,21 @@ def remove_negatives(detections, class_names, num):
     return predictions
 
 
+def remove_negatives_faster(detections, class_names, num):
+    """
+    Faster version of remove_negatives (very useful when using yolo9000)
+    """
+    predictions = []
+    for j in range(num):
+        if detections[j].best_class_idx == -1:
+            continue
+        name = class_names[detections[j].best_class_idx]
+        bbox = detections[j].bbox
+        bbox = (bbox.x, bbox.y, bbox.w, bbox.h)
+        predictions.append((name, detections[j].prob[detections[j].best_class_idx], bbox))
+    return predictions
+
+
 def detect_image(network, class_names, image, thresh=.5, hier_thresh=.5, nms=.45):
     """
         Returns a list with highest confidence class and their bbox

darknet_video.py

@@ -136,7 +136,7 @@ def inference(darknet_image_queue, detections_queue, fps_queue):
 
 def drawing(frame_queue, detections_queue, fps_queue):
     random.seed(3)  # deterministic bbox colors
-    video = set_saved_video(cap, args.out_filename, (darknet_width, darknet_height))
+    video = set_saved_video(cap, args.out_filename, (video_width, video_height))
     while cap.isOpened():
         frame = frame_queue.get()
         detections = detections_queue.get()
@@ -149,7 +149,6 @@ def drawing(frame_queue, detections_queue, fps_queue):
             image = darknet.draw_boxes(detections_adjusted, frame, class_colors)
             if not args.dont_show:
                 cv2.imshow('Inference', image)
-            image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
             if args.out_filename is not None:
                 video.write(image)
             if cv2.waitKey(fps) == 27:
@@ -177,6 +176,8 @@ def drawing(frame_queue, detections_queue, fps_queue):
     darknet_height = darknet.network_height(network)
     input_path = str2int(args.input)
     cap = cv2.VideoCapture(input_path)
+    video_width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
+    video_height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
     Thread(target=video_capture, args=(frame_queue, darknet_image_queue)).start()
     Thread(target=inference, args=(darknet_image_queue, detections_queue, fps_queue)).start()
     Thread(target=drawing, args=(frame_queue, detections_queue, fps_queue)).start()

scripts/deploy-cuda.ps1

@@ -1,18 +1,18 @@
 #!/usr/bin/env pwsh
 
-$url = 'https://developer.download.nvidia.com/compute/cuda/11.3.0/network_installers/cuda_11.3.0_win10_network.exe'
+$url = 'https://developer.download.nvidia.com/compute/cuda/11.4.0/network_installers/cuda_11.4.0_win10_network.exe'
 
-$CudaFeatures = 'nvcc_11.3 cuobjdump_11.3 nvprune_11.3 cupti_11.3 memcheck_11.3 nvdisasm_11.3 nvprof_11.3 ' + `
- 'visual_studio_integration_11.3 visual_profiler_11.3 visual_profiler_11.3 cublas_11.3 cublas_dev_11.3 ' + `
- 'cudart_11.3 cufft_11.3 cufft_dev_11.3 curand_11.3 curand_dev_11.3 cusolver_11.3 cusolver_dev_11.3 ' + `
- 'cusparse_11.3 cusparse_dev_11.3 npp_11.3 npp_dev_11.3 nvrtc_11.3 nvrtc_dev_11.3 nvml_dev_11.3 ' + `
- 'occupancy_calculator_11.3 '
+$CudaFeatures = 'nvcc_11.4 cuobjdump_11.4 nvprune_11.4 cupti_11.4 memcheck_11.4 nvdisasm_11.4 nvprof_11.4 ' + `
+ 'visual_studio_integration_11.4 visual_profiler_11.4 visual_profiler_11.4 cublas_11.4 cublas_dev_11.4 ' + `
+ 'cudart_11.4 cufft_11.4 cufft_dev_11.4 curand_11.4 curand_dev_11.4 cusolver_11.4 cusolver_dev_11.4 ' + `
+ 'cusparse_11.4 cusparse_dev_11.4 npp_11.4 npp_dev_11.4 nvrtc_11.4 nvrtc_dev_11.4 nvml_dev_11.4 ' + `
+ 'occupancy_calculator_11.4 '
 
 try {
   Write-Host 'Downloading CUDA...'
-  Invoke-WebRequest -Uri $url -OutFile "cuda_11.3.0_win10_network.exe"
+  Invoke-WebRequest -Uri $url -OutFile "cuda_11.4.0_win10_network.exe"
   Write-Host 'Installing CUDA...'
-  $proc = Start-Process -PassThru -FilePath "./cuda_11.3.0_win10_network.exe" -ArgumentList @('-s ' + $CudaFeatures)
+  $proc = Start-Process -PassThru -FilePath "./cuda_11.4.0_win10_network.exe" -ArgumentList @('-s ' + $CudaFeatures)
   $proc.WaitForExit()
   $exitCode = $proc.ExitCode
   if ($exitCode -eq 0) {

scripts/deploy-cuda.sh

@@ -7,17 +7,16 @@ else
     sudo apt-get update
     sudo apt-get install build-essential g++
     sudo apt-get install apt-transport-https ca-certificates gnupg software-properties-common wget
-    wget http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/cuda-repo-ubuntu1804_10.2.89-1_amd64.deb
-    sudo apt-key adv --fetch-keys http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/7fa2af80.pub
-    sudo dpkg -i cuda-repo-ubuntu1804_10.2.89-1_amd64.deb
-    wget http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1804/x86_64/nvidia-machine-learning-repo-ubuntu1804_1.0.0-1_amd64.deb
-    sudo dpkg -i nvidia-machine-learning-repo-ubuntu1804_1.0.0-1_amd64.deb
+    sudo wget -O /etc/apt/preferences.d/cuda-repository-pin-600 https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/cuda-ubuntu1804.pin
+    sudo apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/7fa2af80.pub
+    sudo add-apt-repository "deb https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/ /"
+    sudo add-apt-repository "deb http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1804/x86_64/ /"
     sudo apt-get update
     sudo apt-get dist-upgrade -y
-    sudo apt-get install -y --no-install-recommends cuda-compiler-10-2 cuda-libraries-dev-10-2 cuda-driver-dev-10-2 cuda-cudart-dev-10-2 cuda-curand-dev-10-2
-    sudo apt-get install -y --no-install-recommends libcudnn7-dev
+    sudo apt-get install -y --no-install-recommends cuda-compiler-11-4 cuda-libraries-dev-11-4 cuda-driver-dev-11-4 cuda-cudart-dev-11-4
+    sudo apt-get install -y --no-install-recommends libcudnn8-dev
     sudo rm -rf /usr/local/cuda
-    sudo ln -s /usr/local/cuda-10.2 /usr/local/cuda
+    sudo ln -s /usr/local/cuda-11.4 /usr/local/cuda
   elif [[ $(cut -f2 <<< $(lsb_release -r)) == "20.04" ]]; then
     sudo apt-get update
     sudo apt-get install build-essential g++
@@ -28,10 +27,10 @@ else
     sudo add-apt-repository "deb http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu2004/x86_64/ /"
     sudo apt-get update
     sudo apt-get dist-upgrade -y
-    sudo apt-get install -y --no-install-recommends cuda-compiler-11-2 cuda-libraries-dev-11-2 cuda-driver-dev-11-2 cuda-cudart-dev-11-2
+    sudo apt-get install -y --no-install-recommends cuda-compiler-11-4 cuda-libraries-dev-11-4 cuda-driver-dev-11-4 cuda-cudart-dev-11-4
     sudo apt-get install -y --no-install-recommends libcudnn8-dev
     sudo rm -rf /usr/local/cuda
-    sudo ln -s /usr/local/cuda-11.2 /usr/local/cuda
+    sudo ln -s /usr/local/cuda-11.4 /usr/local/cuda
   else
     echo "Unable to deploy CUDA on this Linux version, please wait for a future script update"
   fi

scripts/setup.ps1

@@ -16,9 +16,9 @@ Push-Location $PSScriptRoot
 
 if ($InstallCUDA) {
   & $PSScriptRoot/deploy-cuda.ps1
-  $env:CUDA_PATH="C:\\Program Files\\NVIDIA GPU Computing Toolkit\\CUDA\\v11.3"
-  $env:CUDA_TOOLKIT_ROOT_DIR="C:\\Program Files\\NVIDIA GPU Computing Toolkit\\CUDA\\v11.3"
-  $env:CUDACXX="C:\\Program Files\\NVIDIA GPU Computing Toolkit\\CUDA\\v11.3\\bin\\nvcc.exe"
+  $env:CUDA_PATH="C:\\Program Files\\NVIDIA GPU Computing Toolkit\\CUDA\\v11.4"
+  $env:CUDA_TOOLKIT_ROOT_DIR="C:\\Program Files\\NVIDIA GPU Computing Toolkit\\CUDA\\v11.4"
+  $env:CUDACXX="C:\\Program Files\\NVIDIA GPU Computing Toolkit\\CUDA\\v11.4\\bin\\nvcc.exe"
   $CUDAisAvailable = $true
 }
 else {

scripts/setup.sh

@@ -51,9 +51,9 @@ if [ "$install_tools" = true ] ; then
       sudo apt-get install -y cmake
       sudo apt-get install -y powershell
       if [ "$bypass_driver_installation" = true ] ; then
-        sudo ln -s /usr/local/cuda-10.2/lib64/stubs/libcuda.so /usr/local/cuda-11.2/lib64/stubs/libcuda.so.1
-        sudo ln -s /usr/local/cuda-10.2/lib64/stubs/libcuda.so /usr/local/cuda-11.2/lib64/libcuda.so.1
-        sudo ln -s /usr/local/cuda-10.2/lib64/stubs/libcuda.so /usr/local/cuda-11.2/lib64/libcuda.so
+        sudo ln -s /usr/local/cuda-11.4/lib64/stubs/libcuda.so /usr/local/cuda-11.4/lib64/stubs/libcuda.so.1
+        sudo ln -s /usr/local/cuda-11.4/lib64/stubs/libcuda.so /usr/local/cuda-11.4/lib64/libcuda.so.1
+        sudo ln -s /usr/local/cuda-11.4/lib64/stubs/libcuda.so /usr/local/cuda-11.4/lib64/libcuda.so
       fi
       export PATH=/usr/local/cuda/bin:$PATH
       export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH
@@ -75,9 +75,9 @@ if [ "$install_tools" = true ] ; then
       sudo apt-get install -y cmake
       sudo apt-get install -y powershell
       if [ "$bypass_driver_installation" = true ] ; then
-        sudo ln -s /usr/local/cuda-11.2/lib64/stubs/libcuda.so /usr/local/cuda-11.2/lib64/stubs/libcuda.so.1
-        sudo ln -s /usr/local/cuda-11.2/lib64/stubs/libcuda.so /usr/local/cuda-11.2/lib64/libcuda.so.1
-        sudo ln -s /usr/local/cuda-11.2/lib64/stubs/libcuda.so /usr/local/cuda-11.2/lib64/libcuda.so
+        sudo ln -s /usr/local/cuda-11.4/lib64/stubs/libcuda.so /usr/local/cuda-11.4/lib64/stubs/libcuda.so.1
+        sudo ln -s /usr/local/cuda-11.4/lib64/stubs/libcuda.so /usr/local/cuda-11.4/lib64/libcuda.so.1
+        sudo ln -s /usr/local/cuda-11.4/lib64/stubs/libcuda.so /usr/local/cuda-11.4/lib64/libcuda.so
       fi
       export PATH=/usr/local/cuda/bin:$PATH
       export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH