Closed Surajj4837-th closed 1 year ago
I m testing the model on a KITTI dataset. There are 1946 images in the set. For that, I have created a Test.txt file in which the content is arranged as an image path followed by a corresponding label path. eg:
/media/ADAS1/ADAS_LiDAR/KITTI/Extracted_Data/training/image_2/000001.png /media/ADAS1/ADAS_LiDAR/KITTI/Extracted_Data/training/label_2/000001.txt /media/ADAS1/ADAS_LiDAR/KITTI/Extracted_Data/training/image_2/000002.png /media/ADAS1/ADAS_LiDAR/KITTI/Extracted_Data/training/label_2/000002.txt . ......
While testing, the following line in datasets.py is sshowing all label file paths: self.label_files = img2label_paths(self.img_files)
self.label_files = img2label_paths(self.img_files)
But a few lines down: nf, nm, ne, nc, n = cache.pop('results')
nf, nm, ne, nc, n = cache.pop('results')
where, nm = 1946 (missing number)
Also, the command prompt shows missing labels:
What am I doing wrong?
Found it, the label format was KITTI inside the given files. Converted the KITTI format to COCO format and gave the paths for the same. It worked.
I m testing the model on a KITTI dataset. There are 1946 images in the set. For that, I have created a Test.txt file in which the content is arranged as an image path followed by a corresponding label path. eg:
While testing, the following line in datasets.py is sshowing all label file paths:
self.label_files = img2label_paths(self.img_files)
But a few lines down:
nf, nm, ne, nc, n = cache.pop('results')
where, nm = 1946 (missing number)
Also, the command prompt shows missing labels:
What am I doing wrong?