I convert instances_val2017.json in coco dataset, there are 5000 images in validation set, but I only got 4952 output txt. I am wondering why there's difference between them.(I expect to have 5000 output txt)

https://github.com/SsaRu/convert2Yolo/blob/441ea073f6f3c7b6197677cc05605fbb41181135/voc.py#L16

Thanks for your share! However, there is a problem in the lines I pointed at. You never get the arguments in that voc.py. Just remove the ifs and the script runs fine!

Hi,

First of all, thank you for making this script.

I downloaded the 2014 database from Coco's own site. I ran the script below. Process is starting but a character error message appears at % 3 process.

So how can i resolve this problem ?

python3 example.py --datasets COCO --img_path /coco/images/train2014/ --label /coco/annotations/instances_train2014.json --convert_output_path ./ --img_type ".jpg" --manipast_path ../ --cls_list_file ../data/coco_15c.names

COCO Parsing Result : False, msg : ERROR : 'ascii' codec can't encode character '\u2588' in position 17: ordinal not in range(128), moreInfo : <class 'UnicodeEncodeError'> Format.py 366

Hi ,
Thanks for the effort .
Lets say I am working with VOC dataset.
As far as I understood this repository will help to convert all the annotations in annotations folder to yolo format. But is it possible to convert only a specific category (for eg: person) and seperate images and annotations in different folder ?

Thanks
Rahul

Hi,

I would like to use the coco dataset for YOLO, but only specific classes and not the entire dataset.

Is it possible to do that with convert2Yolo?
I tried to create a file coco.names and I wrote only the classes that interested me, but it didn't worked.
When I used convert2yolo, I got this error :

COCO parsing : 100.0% complete

Yolo Generating Result : False, msg : ERROR : 'hot dog' is not in list, moreInfo : <class 'ValueError'> Format.py

Hot dog is not a class that I write in tje coco.names.

I don't think it should be a requirement that every image has 1 or more labelled objects.

Code works fine (and outputs a 0-length YOLO output) simply by changing Format.py from returning error condition to printing a warning:

                 if len(objects) == 0:
-                    return False, "number object zero"
+                    #return False, "number object zero"
+                    print("WARNING: no objects in file: {}".format(filename))

안녕하세요.
좋은 코드를 올려주셔서 감사합니다.
제가 실행해봤는데 잘 되더라고요.

다만 긍금한 거 하나 있습니다.
제가 사실은 voc datasets의 있는 car, bus, motorbike 클래스에대한 image와 annotation를 따로 저장했습니다.
annotation를 yolo형식으로 바꿀때 3가지 car,bus, motorbike에대한만 나올 수 있도록 voc.names 파일을 "car,bus,motorbike" 3가지만 저정했는데 오류 떴더라고요.
그래서 voc.names파일을 그대로 사용하고 코드를 어떻게 수정해야되나요?

도와주셔서 너무 너무 감사합니다.

I was trying to run your example.py, and gave all relevant arguments to it. This is what I get as an error:

VOC Parsing Result : False, msg : ERROR : 'utf-8' codec can't decode byte 0xff in position 0: invalid start byte, moreInfo : <class 'UnicodeDecodeError'> Format.py 232

Do you have any idea why this could be?
Thanks.

hello
I use voc 2 yolo and i have two classes.
The output of the same image are separated into 2 txt with all 0 and all 1
it canont be used in the yolo training progress
thx!

Hi,

I tried to copy paste the example code from your readme but I found out (from the error messages and help) that you renamed these arguments

--image_path became --img_path
--image_type became --img_type
--clas_list_file became --cls_list_file

I prepared a pull request but Git tells me I don't have access to your repo

When I run the example.py this is the error i get

There are a lot of path problems in your code. For example you should combine the paths with a method, not just with the + operator.

And the output path for the text files is wrong in the convert2Yolo.py. It should be:

result_outpath = str(label_dir + xml_name[:-3] + "txt")

Or for example you are not just reading .xml files.

I would add more flexibility to the images types (at the moment I need to choose one). Images could be .jpg, .png, .jpeg, etc. in the same dataset.

After running the script i got this error msg

YOLO Generating Result : False, msg : ERROR : 'dining table' is not in list, moreInfo : <class 'ValueError'> Format.py 704

any solution ?

Hello,

I get this error when trying to parse VOC. Can somebody help me with that? Thanks

grep -iE "manifast|manipast" *py

Is this meant to be manifest?

There are nearly 240MB inside .git/objects. Cloud you please clear them all otherwise we can't use git clone directly since it will download all include this giant file.

Thanks~

I have this issue with following line of code.
python example.py --datasets VOC --img_path invoices-PascalVOC-export/JPEGImages --label invoices-PascalVOC-export/Annotations --convert_output_path yolo_converted --img_type ".jpg" --manipast_path yolo_converted --cls_list_file C:\Users\fkhalil\primeStone\deep_net\darknet\data\invoiceLabels\invoices-PascalVOC-export\classes.names

Error message:
VOC Parsing Result : False, msg : ERROR : , moreInfo : <class 'StopIteration'>  Format.py       224

Note: Classes contains the name of all classed I have use in Vott application to label the images.

Hello,
I am trying to convert over 500,000 files and each little error stops the processing and I have to then hunt for the missing item in my xml files. For example, it appears that one or more of my xml files were missing the tag, so instead of identifying the file(s), the script throws and error and quits.
In another very common error, the tag is missing ion one or more files
Another error type: The value in the classes file does not match the name value in the xml file. This is very common, in case of a typo. For example, one of my files could have a misspelling. But the system stops processing.

Ideally, since it is very common to process tens of thousands of files, a better way of handling this would be to ( a ) print the name of the offending file on the screen ( b ) or better still, rename the offending file, by adding a suffix, so that one can go through all invalid files and fix them.

Hope you'll take my suggestion.

Right now, I'm processing 500K files and am on my 15th iteration, as it stops after every single error
Thanks

I'm converting this dataset (https://www.di.ens.fr/willow/research/headdetection/) from xml to txt to use it in YOLO but I'm getting this error.
VOC Parsing Result : False, msg : ERROR : 'NoneType' object has no attribute 'text', moreInfo : <class 'AttributeError'> Format.py 258

this is line 258 in Format.py
tmp = {
"name": _object.find("name").text
}

Do you know how to solve it?

I saw that in a previous github issue, 6 classes had to be added to the 80 existing classes in the coco.names list (though they are only rewordings of objects already in the dataset):

motorcycle
potted plant
dining table
tv
couch
airplane

This causes some of the .txt annotation files to have classes numbered up to 86, which causes errors when I try to run known implementations of YOLO (like the AlexeyAB repository, which expects classes only up to 80). Can somebody please explain?

Hello,
Thank you very much for the effot you put into this code but I tried to run this code on python 2.7 so I wrote the command: python example.py --datasets VOC --img_path ~/VOC/Images/ --label ~/VOC/Annotations/ --convert_output_path ~/VOC/YOLO/ --img_type ".jpg" --manipast_path ./ --cls_list_file ~/VOC/voc.names and got the following syntax error:
Traceback (most recent call last):
File "example.py", line 11, in
from Format import VOC, COCO, UDACITY, KITTI, YOLO
File "/home/abanoub/Desktop/convert2Yolo-master/Format.py", line 71
print('\r%s|%s| %s%% (%s/%s) %s' % (prefix, bar, percent, iteration, total, suffix), end = '\r')
^
SyntaxError: invalid syntax

Do know what would be a solution to such a problem?
Cheers!

@raggot
How did you fix this problem:
python3 example.py --datasets KITTI --img_path /KITTI/images_new --label /KITTI/labels --convert_output_path ~/kitti --img_type ".jpg" --cls_list_file KITTI/my-classes.names
KITTI Parsing Result : False, msg : ERROR : , moreInfo : <class 'StopIteration'> Format.py 484

Please share. Thank You

Originally posted by @lionverve2015 in #6 (comment)

Hi,

I have the following issue, do you know where the issue is?
VOC Parsing Result : False, msg : ERROR : 'utf-8' codec can't decode byte 0xb0 in position 37: invalid start byte, moreInfo : <class 'UnicodeDecodeError'> Format.py 233

I am using the VOTT Tool to label my pictures. With the VOTT I am exporting the labels to Pascal VOC in order to convert it to Yolo.

Can you help please?

Thank you very much!
Khani

Hello @ssaru ,

I'm trying to execute the script like this:
python example.py --datasets KITTI --img_path ./images/000021.jpg --label ./images/000021.txt --convert_output_path test.txt --img_type ".jpg" --manipast_path manipast.txt --cls_list_file names.txt

But it doesn't work. I don't understand what is manipast_path.
Would it be possible to add a more specific example where all the files are present in the git repository?

Thank you.

If for any reason an XML file would not have "object" entity, the script chokes at

I'd suggest to close the xml"file before removing it from them "xml_path".

Thanks.