I was following along with the YOLOS tutorial, and I saw input images are passed to feature_extractor in the init method.
The actual model seems to accept actual images with 3 channels, (the model attribute of the Detr class)
what does the feature_extractor do?
All the object detection models seem to use the feature_extractor,
is it something that has to do with transformer?
I was following along with the YOLOS tutorial, and I saw input images are passed to feature_extractor in the init method. The actual model seems to accept actual images with 3 channels, (the model attribute of the Detr class) what does the feature_extractor do?
All the object detection models seem to use the feature_extractor, is it something that has to do with transformer?