the question about speed of detecting

[zhuozhudd]

Hi ! I have a question when I learn your codes recently about evaluation.

INFO:root:CUDA detected. Running with GPU acceleration. INFO:root:Loading Cornell Dataset... INFO:root:Validation size: 89 INFO:root:Done INFO:root: Evaluating model iou_0.97 INFO:root:Average evaluation time per image: 134.32957349198588ms

L:\WS\gr-convnet>python evaluate.py --dataset cornell --dataset-path L:/WS/Cornell --network iou_0.97 --iou-eval INFO:root:CUDA detected. Running with GPU acceleration. INFO:root:Loading Cornell Dataset... INFO:root:Validation size: 89 INFO:root:Done INFO:root: Evaluating model iou_0.97 INFO:root:Average evaluation time per image: 138.0717566843783ms INFO:root:IOU Results: 85/89 = 0.955056

L:\WS\gr-convnet>python evaluate.py --dataset cornell --dataset-path L:/WS/Cornell --network iou_0.97 --iou-eval INFO:root:CUDA detected. Running with GPU acceleration. INFO:root:Loading Cornell Dataset... INFO:root:Validation size: 89 INFO:root:Done INFO:root: Evaluating model iou_0.97 INFO:root:Average evaluation time per image: 136.26244630706444ms INFO:root:IOU Results: 85/89 = 0.955056

Some logs above when I run evaluate.py. I used your trained-models to evaluate. I got 136ms but in your paper it is only 20ms.How can I got this result.

Thanks.

[skumra]

The runtime speed is a function of the hardware being used for computation. The results reported in the paper are for the hardware we used for the experiments. You can refer to the paper for the hardware setup used for the experiments.

[praj441]

I think the speed mentioned in the paper should correspond to the batch_wise processing of the test set. When I run the inference with a single image, I also get around 140ms. Regarding the hardware setup, I am having a similarly powerful machine than is mentioned in the paper. In the paper - Intel Core i7-7800X CPU clocked at 3.50 GHz and an NVIDIA GeForce GTX 1080 Ti Mine is 16 core, Intel(R) Xeon(R) CPU E5-2620 v4 @ 2.10GHz and 3 NVIDIA GeForce GTX 1080 Ti

[praj441]

Let me

I think the speed mentioned in the paper should correspond to the batch_wise processing of the test set. When I run the inference with a single image, I also get around 140ms. Regarding the hardware setup, I am having a similarly powerful machine than is mentioned in the paper. In the paper - Intel Core i7-7800X CPU clocked at 3.50 GHz and an NVIDIA GeForce GTX 1080 Ti Mine is 16 core, Intel(R) Xeon(R) CPU E5-2620 v4 @ 2.10GHz and 3 NVIDIA GeForce GTX 1080 Ti

Let me correct myself. The time mentioned in the paper is correct. I am able to get the 20ms speed over a single image processing. The problem is strange to me. When I pass the first sample, it takes around 150ms. But the subsequent samples (in a for loop), it's processing faster. The same is happening during the training (not just this repo code but every other code I tried so far), the first forward pass always takes a much longer time. Does anyone know the reason behind this?