Square training and Rectangular inference?

[github2016-yuan]

❔Question

I use this repo to train my custom dataset and it works well but I still have some questions here.

Resolution of image in my dataset is 3840 * 1080 and I use cmd like below to train: (train.py)
--img 640 --batch 16 --epochs 300 --weights yolov5s.pt

It means I use square training (image is scaled into 640 * 640 with padding) and I get best.pt and last.pt sucessfully.

Then I use cmd below to detect: (detect.py)

--source a.mp4

I get expected nice result.

It means I use square inference but I find something odd while debug in detect.py

the actual resolution of detected image is scaled to 640 * 192. That is to say I use rectangular inference here in fact.

In my opinion, training and inference should have the same resolution but here it is not.

Glad to get some advice.

[github-actions]

👋 Hello @github2016-yuan, thank you for your interest in 🚀 YOLOv5! Please visit our ⭐️ Tutorials to get started, where you can find quickstart guides for simple tasks like Custom Data Training all the way to advanced concepts like Hyperparameter Evolution.

If this is a 🐛 Bug Report, please provide screenshots and minimum viable code to reproduce your issue, otherwise we can not help you.

If this is a custom training ❓ Question, please provide as much information as possible, including dataset images, training logs, screenshots, and a public link to online W&B logging if available.

For business inquiries or professional support requests please visit https://www.ultralytics.com or email Glenn Jocher at glenn.jocher@ultralytics.com.

Requirements

Python 3.8 or later with all requirements.txt dependencies installed, including torch>=1.7. To install run:

$ pip install -r requirements.txt

Environments

YOLOv5 may be run in any of the following up-to-date verified environments (with all dependencies including CUDA/CUDNN, Python and PyTorch preinstalled):

• Google Colab and Kaggle notebooks with free GPU:
• Google Cloud Deep Learning VM. See GCP Quickstart Guide
• Amazon Deep Learning AMI. See AWS Quickstart Guide
• Docker Image. See Docker Quickstart Guide

Status

If this badge is green, all YOLOv5 GitHub Actions Continuous Integration (CI) tests are currently passing. CI tests verify correct operation of YOLOv5 training (train.py), testing (test.py), inference (detect.py) and export (export.py) on MacOS, Windows, and Ubuntu every 24 hours and on every commit.

[glenn-jocher]

@github2016-yuan 👋 Hello, thank you for asking about the differences between train.py, detect.py and test.py in YOLOv5.

These 3 files are designed for different purposes and utilize different dataloaders with different settings. train.py dataloaders are designed for a speed-accuracy compromise, test.py is designed to obtain the best mAP on a validation dataset, and detect.py is designed for best real-world inference results. A few important aspects of each:

train.py

• trainloader LoadImagesAndLabels(): designed to load train dataset images and labels. Augmentation capable and enabled.

  yolov5/train.py

  Lines 188 to 192 in fca5e2a

  	# Trainloader
  	dataloader, dataset = create_dataloader(train_path, imgsz, batch_size, gs, opt,
  	hyp=hyp, augment=True, cache=opt.cache_images, rect=opt.rect, rank=rank,
  	world_size=opt.world_size, workers=opt.workers,
  	image_weights=opt.image_weights, quad=opt.quad, prefix=colorstr('train: '))

• testloader LoadImagesAndLabels(): designed to load val dataset images and labels. Augmentation capable but disabled.

  yolov5/train.py

  Lines 199 to 202 in fca5e2a

  	testloader = create_dataloader(test_path, imgsz_test, batch_size * 2, gs, opt, # testloader
  	hyp=hyp, cache=opt.cache_images and not opt.notest, rect=True, rank=-1,
  	world_size=opt.world_size, workers=opt.workers,
  	pad=0.5, prefix=colorstr('val: '))[0]

• image size: 640
• rectangular inference: False
• confidence threshold: 0.001
• iou threshold: 0.6
• multi-label: True
• padding: None

test.py

• dataloader LoadImagesAndLabels(): designed to load train, val, test dataset images and labels. Augmentation capable but disabled.

  yolov5/test.py

  Lines 89 to 90 in fca5e2a

  	dataloader = create_dataloader(data[task], imgsz, batch_size, gs, opt, pad=0.5, rect=True,
  	prefix=colorstr(f'{task}: '))[0]

• image size: 640
• rectangular inference: True
• confidence threshold: 0.001
• iou threshold: 0.6
• multi-label: True
• padding: 0.5 * maximum stride

detect.py

• dataloaders (multiple): designed for loading multiple types of media (images, videos, globs, directories, streams).

  yolov5/detect.py

  Lines 46 to 53 in fca5e2a

  	# Set Dataloader
  	vid_path, vid_writer = None, None
  	if webcam:
  	view_img = check_imshow()
  	cudnn.benchmark = True # set True to speed up constant image size inference
  	dataset = LoadStreams(source, img_size=imgsz, stride=stride)
  	else:
  	dataset = LoadImages(source, img_size=imgsz, stride=stride)

• image size: 640
• rectangular inference: True
• confidence threshold: 0.25
• iou threshold: 0.45
• multi-label: False
• padding: None

YOLOv5 PyTorch Hub

models.autoShape() class used for image loading, preprocessing, inference and NMS. For more info see YOLOv5 PyTorch Hub Tutorial

yolov5/models/common.py

Lines 225 to 250 in fca5e2a

	class autoShape(nn.Module):
	# input-robust model wrapper for passing cv2/np/PIL/torch inputs. Includes preprocessing, inference and NMS
	conf = 0.25 # NMS confidence threshold
	iou = 0.45 # NMS IoU threshold
	classes = None # (optional list) filter by class
	def __init__(self, model):
	super(autoShape, self).__init__()
	self.model = model.eval()
	def autoshape(self):
	print('autoShape already enabled, skipping... ') # model already converted to model.autoshape()
	return self
	@torch.no_grad()
	@torch.cuda.amp.autocast(torch.cuda.is_available())
	def forward(self, imgs, size=640, augment=False, profile=False):
	# Inference from various sources. For height=640, width=1280, RGB images example inputs are:
	# filename: imgs = 'data/samples/zidane.jpg'
	# URI: = 'https://github.com/ultralytics/yolov5/releases/download/v1.0/zidane.jpg'
	# OpenCV: = cv2.imread('image.jpg')[:,:,::-1] # HWC BGR to RGB x(640,1280,3)
	# PIL: = Image.open('image.jpg') # HWC x(640,1280,3)
	# numpy: = np.zeros((640,1280,3)) # HWC
	# torch: = torch.zeros(16,3,320,640) # BCHW (scaled to size=640, 0-1 values)
	# multiple: = [Image.open('image1.jpg'), Image.open('image2.jpg'), ...] # list of images

• image size: 640
• rectangular inference: True
• confidence threshold: 0.25
• iou threshold: 0.45
• multi-label: False
• padding: None

[fabiozappo]

@github2016-yuan this could help you understanding the benefits of rectangular inference in detect.py: ultralytics/yolov3#232

[github2016-yuan]

Fine @fabiozappo