YOLO-Annotation-Tool
YOLO-Annotation-Tool copied to clipboard
ManivannanMurugavel
invalid literal for float()
I'm trying to train my own data on the Yolo network, but before that I have to convert the bounding boxes co-ordinates to the form it wants. and i get the error: first it prints out all the file names and the content of the data then
1 29 45 615 251
['1\n29', '45', '615', '251\n']
1
29
(803, 624)
Traceback (most recent call last):
File "convert.py", line 96, in
@ManivannanMurugavel I have the same issue, which python version should I use?
Hello dear, thanks for your response, in the tool you pre-define two classes while labeling images, however i am working on single class problem, i am detecting a screen from the given images, How can i manage it? can i remove classes from the code? @ManivannanMurugavel
Now i am using new tool, i change the classes in the Classes.txt file, then i labeled 10 images to test it and after successfully labeling them i run process. py, there was no error buy test.txt and train.txt files are empty.
This is not an issue. Please check the directory which contains the all .txt files.
Bro now i have 10 sample images in the Images/001 folder and 10 corresponding .txt files in the Labels/001 folder. So how i can start training Yolo2 or Yolo3 model ? are these images already in the Yolo formate? How can i create train.txt and test.txt files or it do not require? @ManivannanMurugavel
Hi @MuhammadAsadJaved, First, you copy all images and it's corresponding .txt files and paste into sample folder. Set your folder path in this line https://github.com/ManivannanMurugavel/YOLO-Annotation-Tool/blob/master/process.py#L8 and run it.
Thank you very much bro, This problem is resolved now. I sent you a message on Faccebook, Please reply if it is appropriate. https://www.facebook.com/M.asad.javed.gujjar @ManivannanMurugavel