Stéphane Charette

Annotations have nothing to do with the size of the images. All YOLO annotations are normalized. You can resize your images to 9999x9999 or 160x160, and the annotations will still...

This is a repeat of the same question you had on the old AlexeyAB repo. :( As I wrote to you there, I do this often, both Windows and Linux....

No new comment in 4 months. Closing. If you have more information, please re-open with the necessary details.

You should use the new Darknet repo, which has updated python support: https://github.com/hank-ai/darknet#table-of-contents

Seems to be working for me: 

Also note that YOLOv7 is probably not what you want to use: https://www.ccoderun.ca/programming/yolo_faq/#configuration_template

> YOLOv7-tiny (shown on your screenshot) works for me too but not YOLOv7. > > I don't know if the tiny models will suffice for COCO-style 80-classes annotations in the...

At the very least, can you confirm your MD5 for the relevant files? > md5sum yolov7* ../cfg/yolov7* 70be176996efd72c37a0e481ca169a02 yolov7-tiny.weights 028a36c8e82aa3b9c41976b37cd4b08d yolov7.weights 484b572c0573bdc2c84c56549fe61481 ../cfg/yolov7-tiny.cfg eb5a8f5b19c6a36c944921a3d1acc7c7 ../cfg/yolov7.cfg 7572052eba31b9f2bd5ddfd1dff0e894 ../cfg/yolov7x.cfg

Run it with the same image as what I used, the ones in the "artwork" directory.

Can you run the same command I ran? I have no idea what "annotate_it_all.names" contains. darknet_02_display_annotated_images ../cfg/coco.names ../cfg/yolov7.cfg yolov7.weights ../artwork/dog.jpg The next thing I'd say is try and run it...