If I run the code for just 1 image, inference time per image is: 0.378s
If I run the code for 10 images, inference time per image is 0.09s
If I increase the count of test image, inference time per image reduces more.
What can I do to get correct inference time for just one image. Which configs should I change to what?
Note: I changed inside imageSets/test.txt to only one image.
Hi,
running code :
python tools/plain_train_net.py --eval-only --config-file "configs/smoke_gn_vector.yaml" --ckpt model_final.pth
If I run the code for just 1 image, inference time per image is: 0.378s If I run the code for 10 images, inference time per image is 0.09s If I increase the count of test image, inference time per image reduces more.
What can I do to get correct inference time for just one image. Which configs should I change to what?
Note: I changed inside imageSets/test.txt to only one image.
Best.
Edit: Sorry my bad. Inference time is ~40ms.