About the performance

[YP-Liu]

Hi, Luca and Jack. good work as always. I run the CFNet-conv2, CFNet-conv5, Baseline-conv5 on OTB2013 OPE by setting the hyperparameters according to the table2 in the paper and find that the results are much worser(the numbers after the "/" are the results shown in the paper):

1. Baseline-conv5: precision:0.779 / 0.806 success:0.6 / 0.618

2. CFNet-conv2: precision:0.737 / 0.807 success:0.561 / 0.611

3. CFNet-conv5: precision:0.723 / 0.803 success:0.518 / 0.611

Does anyone knows what may cause these results? I run the code on Matlab2016a by using CUDA8.0 and matconvnet-25

[lianghao12138]

I also met this problem.

[YP-Liu]

Do you get similar results as i do?

[lianghao12138]

I run SIamFC(3s) on OTB2013 and got the following results: precision: 0.762 success: 0.580

[bertinetto]

I think we have some problems with newer versions of matconvnet/matlab/cuda/cudnn :
Soon I will try to get the exact versions we use and report them on README, but I think we used Matlab 2015b, CUDA7.5, cudnn 5.1, matconvnet-23.

[YP-Liu]

I try to generate random hyperparameters to run baseline-conv5 and cfnet-conv2 and find that the hyperparameters associated to best performance seem to have a huge difference from those given in the paper.For example, I get better performance(0.805/0.618) by setting hyperparameters in baseline-conv5 as blow: scaleStep: 1.0355 scalePenalty: 0.9815 scaleLR: 0.575 zLR: 0.006 wInfluence: 0.225

[ArsNova123]

so,can you give all performance and the corresponding hyperparameters to me?

[Lauretta1995]

I think we have some problems with newer versions of matconvnet/matlab/cuda/cudnn : Soon I will try to get the exact versions we use and report them on README, but I think we used Matlab 2015b, CUDA7.5, cudnn 5.1, matconvnet-23.

I use Matlab 2015b, CUDA8.0, cudnn 6.02, matconvnet-24, the OPE of OTB100 is 52.1(Iou) and 66.7(prec) based on Baseline+CF-conv3, can you give me a suggestion? I should change the version cuda or cudnn or matconvnet? @bertinetto