You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I created a new project and a task and uploaded images I want to label (bounding box). When I try to upload annotations, I get the following error:
"Could not upload annotation for the cvat.apps.dataset_manager.bindings.CvatDatasetNotFoundError: Check format docs. Dataset must contain a file: "obj.data"
I thought the issue was with the paths for the labels and images. I tried making a new project, uploading 10 images, annotating them, and exporting the results. I then copied the folder paths exactly as they were in the exported folder (e.g. project_name/obj_train_data/image_0000.txt). Within the project_name folder, there are files called obj.data, obj.names, train.txt. I have then compressed the project_name folder and tried to upload that.
I have tried many variations on this (e.g. different folder paths, adding a dataset.yaml, including the images folder when I upload the annotations) but I always get the same error. I have included a log of the error.
I am also including the zip file I used to upload the annotations.
Dataset size: 8786 images
The dataset has already been annotated but I know there are errors in some of the bounding boxes. I am not sure if these errors (e.g. negative numbers) would cause the issues, but I can view those images and bounding boxes when matplotlib.
Environment
OS: MacOS Sonoma 14.7
Browser: Google Chrome
CVAT Server version: 2.23.2
CVAT UI version: 1.67.0
The text was updated successfully, but these errors were encountered:
Unzipping on Mac extracting correct structure. So, this is Mac only related issue.
Anyway, I believe we should add support for such mac os created archives in CVAT.
Actions before raising this issue
Steps to Reproduce
I created a new project and a task and uploaded images I want to label (bounding box). When I try to upload annotations, I get the following error:
"Could not upload annotation for the cvat.apps.dataset_manager.bindings.CvatDatasetNotFoundError: Check format docs. Dataset must contain a file: "obj.data"
I thought the issue was with the paths for the labels and images. I tried making a new project, uploading 10 images, annotating them, and exporting the results. I then copied the folder paths exactly as they were in the exported folder (e.g. project_name/obj_train_data/image_0000.txt). Within the project_name folder, there are files called obj.data, obj.names, train.txt. I have then compressed the project_name folder and tried to upload that.
I have tried many variations on this (e.g. different folder paths, adding a dataset.yaml, including the images folder when I upload the annotations) but I always get the same error. I have included a log of the error.
I am also including the zip file I used to upload the annotations.
testing.zip
cvat_error_log.txt
Expected Behavior
No response
Possible Solution
No response
Context
Dataset size: 8786 images
The dataset has already been annotated but I know there are errors in some of the bounding boxes. I am not sure if these errors (e.g. negative numbers) would cause the issues, but I can view those images and bounding boxes when matplotlib.
Environment
The text was updated successfully, but these errors were encountered: