Hi, I am trying to find the images that are not detected during the evaluation of model with TAO , Is there any method for this one in Tao.
May I know which network you are running?
Hi Morganh, I am running Dino and Mask_Rcnn.
For “the images that are not detected” in Dino evaluation, do you mean there is not any FP or TP on the images?
Hi Morganh, suppose if i get mAp=0.92 then i want to know the images where the model is making mistakes to detect accurately. Simply i want too find the complex images that model feels difficult to detect and causing low mAP.
OK, you can run
tao model dino inference xxx to get the output images to check which are playing not well.
Yes Morganh, I already did that just checking with you if there is any method like producing log file during evaluation to see the images that model feels difficult to detect.
Not evaluation. You can run
tao model dino inference . The
inference tool produces annotated image outputs and txt files that contain prediction information.
Got it Morganh, Thanks a lot.
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.