PyTorch-YOLOv3
PyTorch-YOLOv3 copied to clipboard
eriklindernoren
how to detect video
I have the same question! I have try to add the part of testing local video or webcam video code , but I meet some problems which I can't solve it.
The problem is 'RuntimeError: Expected 4-dimensional input for 4-dimensional weight [32, 3, 3, 3], but got 3-dimensional input of size [480, 640, 3] instead' when I try to transform the frame image to tensor, but it occurs the error above.
I have the same question! I have try to add the part of testing local video or webcam video code , but I meet some problems which I can't solve it. I also tried to add the test code, but because the weight problem could not solve the problem of the test video, I finally replaced someone else's yolov3 GitHub
I think the way to modify the code is wrong, I also struggling for it.
I also find other yolov3 on Github, but it doesn't show in good quality.
Could you tell me what yolov3 you are using? Thanks a lot
git clone https://github.com/ultralytics/yolov3.git
I appreciate you can share this one. And I have a question, Is it you just replaced the file about video detect?
I appreciate you can share this one. And I have a question, Is it you just replaced the file about video detect?
I think you may have misunderstood what I mean, because I couldn't test the video, so I directly changed yolov3 and used the code of the link blogger above, which contains the command to test the video directly.
from future import division
from models import * from utils.utils import * from utils.datasets import *
import os import sys import time import datetime import argparse import cv2
from PIL import Image
import torch from torch.utils.data import DataLoader from torchvision import datasets from torch.autograd import Variable
import matplotlib.pyplot as plt import matplotlib.patches as patches from matplotlib.ticker import NullLocator
if name == "main": parser = argparse.ArgumentParser() parser.add_argument("--image_folder", type=str, default="data/samples", help="path to dataset") parser.add_argument("--vedio_file", type=str, default="vedio_samples/2.mp4", help="path to dataset") parser.add_argument("--model_def", type=str, default="config/yolov3-tiny.cfg", help="path to model definition file") parser.add_argument("--weights_path", type=str, default="model_trained/100-epoch-air.pth", help="path to weights file") parser.add_argument("--class_path", type=str, default="data/air.names", help="path to class label file") parser.add_argument("--conf_thres", type=float, default=0.8, help="object confidence threshold") parser.add_argument("--nms_thres", type=float, default=0.4, help="iou thresshold for non-maximum suppression") parser.add_argument("--batch_size", type=int, default=1, help="size of the batches") parser.add_argument("--n_cpu", type=int, default=3, help="number of cpu threads to use during batch generation") parser.add_argument("--img_size", type=int, default=416, help="size of each image dimension") parser.add_argument("--checkpoint_model", type=str, help="path to checkpoint model") opt = parser.parse_args() print(opt) device = torch.device("cuda" if torch.cuda.is_available() else "cpu") model = Darknet(opt.model_def, img_size=opt.img_size).to(device) if opt.weights_path.endswith(".weights"): # Load darknet weights model.load_darknet_weights(opt.weights_path) else: # Load checkpoint weights model.load_state_dict(torch.load(opt.weights_path)) model.cuda() model.eval() # Set in evaluation mode classes = load_classes(opt.class_path) Tensor = torch.cuda.FloatTensor if torch.cuda.is_available() else torch.FloatTensor if opt.vedio_file.endswith(".mp4"): cap = cv2.VideoCapture(opt.vedio_file) colors = np.random.randint(0, 255, size=(len(classes), 3), dtype="uint8") a=[] while cap.isOpened(): ret, img = cap.read() PILimg = np.array(Image.fromarray(cv2.cvtColor(img,cv2.COLOR_BGR2RGB))) imgTensor = transforms.ToTensor()(PILimg) imgTensor, _ = pad_to_square(imgTensor, 0) imgTensor = resize(imgTensor, 416) imgTensor = imgTensor.unsqueeze(0) imgTensor = Variable(imgTensor.type(Tensor))
with torch.no_grad():
detections = model(imgTensor)
detections = non_max_suppression(detections, opt.conf_thres, opt.nms_thres)
a.clear()
if detections is not None:
a.extend(detections)
b=len(a)
if len(a) :
for detections in a:
if detections is not None:
detections = rescale_boxes(detections, opt.img_size, PILimg.shape[:2])
unique_labels = detections[:, -1].cpu().unique()
n_cls_preds = len(unique_labels)
for x1, y1, x2, y2, conf, cls_conf, cls_pred in detections:
box_w = x2 - x1
box_h = y2 - y1
color = [int(c) for c in colors[int(cls_pred)]]
print(cls_conf)
img = cv2.rectangle(img, (x1, y1 + box_h), (x2, y1), color, 2)
cv2.putText(img, classes[int(cls_pred)], (x1, y1), cv2.FONT_HERSHEY_SIMPLEX, 0.5, color, 2)
cv2.putText(img, str("%.2f" % float(conf)), (x2, y2 - box_h), cv2.FONT_HERSHEY_SIMPLEX, 0.5,
color, 2)
print()
print()
#cv2.putText(img,"Hello World!",(400,50),cv2.FONT_HERSHEY_PLAIN,2.0,(0,0,255),2)
cv2.imshow('frame', img)
#cv2.waitKey(0)
if cv2.waitKey(25) & 0xFF == ord('q'):
break
cap.release()
cv2.destroyAllWindows()
I write a simple vedio base on this
@Guardian-Li It looks like to be similar to my code, but how do you solve the problem called:
RuntimeError: Expected 4-dimensional input for 4-dimensional weight [32, 3, 3, 3], but got 3-dimensional input of size [480, 640, 3] instead?
@Guardian-Li almost the same code the difference is that I haven't resized the frame from the camera
the Tensor you change from image must be use imgTensor = imgTensor.unsqueeze(0)
@Guardian-Li @PommesPeter Did you solve it? I need the detect.py file for this repo to detect video.
@Guardian-Li @PommesPeter is there a working detect.py that works for video? The python code posted above by @Guardian-Li, is that working code?
@aditjha it works, I follow his/her code and my understanding of the code, and i make it.
@impravin22 Yes, I solve this problem. I follow his/her(@Guardian-Li ) code and I make it
@PommesPeter Great. Can you please send your detect.py if you dont mind.
Thanks.
@PommesPeter thank you for replying! Also, I am new to all this...using yolo for a project...so this detect.py allows for a recorded video to be inferenced correct?
@PommesPeter Yeah I found it. It is as video.py in @Guardian-Li repo, isn't it?
Thank you very much