About TP, TN, FP, FN

[ferro07]

❔Question

Hi @glenn-jocher
Thanks for this great work.
I am trying to add Tp, Fp, Fn to the test results. Firstly, I am using this class for calling only tp : def ap_per_class(tp, conf, pred_cls, target_cls, plot=False, fname='precision-recall_curve.png'):

On the test.py, adding tp to the # Compute statistics, but I am getting followings

raceback (most recent call last):
File "test.py", line 272, in
test(opt.data,
File "test.py", line 88, in test
s = ('%20s' + '%12s' * 6) % ('Class', 'Images', 'Targets', 'P', 'R', 'mAP@.5', 'mAP@.5:.95', 'tp')
TypeError: not all arguments converted during string formatting

Thanks in advance for your help

Additional context

[glenn-jocher]

@ferro07 you need to put breakpoints near these variables to understand what you're getting into, as these are large multidimensional matrices of 1's and 0's. I don't know what you want to do with them exactly but I doubt printing them to screen would be very useful.

[jaqub-manuel]

@glenn-jocher thanks for quick reply.
I also want to find the number of FNs and thus, I can try to apply different techniques to reduce FP, FN or increase TP so on.
(Precision = TP/TP+FP and Recall = TP/TP+FN), if I print this at test.py, I will analyze my custom dataset in more detail.
You have this values in general.py (def ap_per_class(tp, conf, pred_cls, target_cls, plot=False, fname='precision-recall_curve.png'):
but, I got errors.
Thanks

[glenn-jocher]

@jaqub-manuel sure. Any code modifications are up to you.

[ZwNSW]

@glenn-jocher I want to talk about tp, fp, fn output to the screen, but I don't know how to modify the code. This is a difficult problem.
Thanks!

[ZwNSW]

@ferro07 Hi,Do you have solved this problem?I want to get the tp,fp,fn to analyze my problom,but I can not print them out to my terminal.

[glenn-jocher]

@ZwNSW TP, FP are computed here:

yolov5/utils/general.py

Lines 250 to 263 in c8c5ef3

	def ap_per_class(tp, conf, pred_cls, target_cls, plot=False, fname='precision-recall_curve.png'):
	""" Compute the average precision, given the recall and precision curves.
	Source: https://github.com/rafaelpadilla/Object-Detection-Metrics.
	# Arguments
	tp: True positives (nparray, nx1 or nx10).
	conf: Objectness value from 0-1 (nparray).
	pred_cls: Predicted object classes (nparray).
	target_cls: True object classes (nparray).
	plot: Plot precision-recall curve at mAP@0.5
	fname: Plot filename
	# Returns
	The average precision as computed in py-faster-rcnn.
	"""

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[glenn-jocher]

@ferro07 @jaqub-manuel good news 😃! Your original issue may now be fixed ✅ in PR #5727. This PR explicitly computes TP and FP from the existing Labels, P, and R metrics:

TP = Recall * Labels
FP = TP / Precision - TP

These TP and FP per-class vectors are left in val.py for users to access if they want:

yolov5/val.py

Line 240 in 36d12a5

tp, fp, p, r, f1, ap, ap_class = ap_per_class(*stats, plot=plots, save_dir=save_dir, names=names)

To receive this update:

• Git – git pull from within your yolov5/ directory or git clone https://github.com/ultralytics/yolov5 again
• PyTorch Hub – Force-reload model = torch.hub.load('ultralytics/yolov5', 'yolov5s', force_reload=True)
• Notebooks – View updated notebooks
• Docker – sudo docker pull ultralytics/yolov5:latest to update your image

Thank you for spotting this issue and informing us of the problem. Please let us know if this update resolves the issue for you, and feel free to inform us of any other issues you discover or feature requests that come to mind. Happy trainings with YOLOv5 🚀!