evaluation code

[pribadihcr]

Hi @bertinetto

Is there example code to measure the performance ?. thanks

[xpain1]

Hi @bertinetto
What's the meanning of '/path/to/deep-one-shot/util'in startup.m?

[bertinetto]

It's a template/example path, you should fill in the absolute path of your
deep-one-shot/util folder

On Thu, Nov 10, 2016 at 8:34 AM, xpain1 notifications@github.com wrote:

Hi @bertinetto https://github.com/bertinetto

What's the meanning of '/path/to/deep-one-shot/util'in startup.m?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#4 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/ADLRueVK2ME7CRYL3BkdKow8uQUUS5Gaks5q8tcXgaJpZM4KIXJT
.

[xpain1]

@bertinetto
Thanks,but I don't find a folder named 'deep-one-shot'?

[bertinetto]

oh i see, that was the name of the private repo.
it's siamese-fc/util

On Thu, Nov 10, 2016 at 10:27 AM, xpain1 notifications@github.com wrote:

@bertinetto https://github.com/bertinetto
Thanks,but I don't find a folder named 'deep-one-shot'?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#4 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/ADLRudOEJGHsWe1HN0EtZT9THUMU3aoJks5q8vGIgaJpZM4KIXJT
.

[xpain1]

@bertinetto
oh i see ,thanks!

[ygren]

Hi，Where can i get your evaluation code in OTB and VOT？I want to reappear your results，Thanks.

[bertinetto]

Hi, I didn't commit the evaluation code but it should be straightforward.
Consider that I have used the network trained with 20% grayscale to the on
OTB (shared in the project page)
Otherwise, I have shared all computed bounding boxes on the project page
http://www.robots.ox.ac.uk/~luca/siamese-fc.html

Cheers,
Luca

On Sat, Nov 12, 2016 at 1:05 PM, YgRen notifications@github.com wrote:

Hi，Where i can get your evaluation code in OTB and VOT，I want to reappear
your results，Thanks.

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
#4 (comment),
or mute the thread
https://github.com/notifications/unsubscribe-auth/ADLRufEySKmqOmRokY5UijZCfqhv2f8zks5q9bmIgaJpZM4KIXJT
.

[xpain1]

Hi, @bertinetto
i met some trouble when i try to train the net .there is non-existent field named 'dilate' in the layer 'conv1',but you use it in the info = vl_simplenn_display(net) ,so i want to know the reason,thanks!
here is my error:
Reference to non-existent field 'dilate'.

Error in vl_simplenn_display (line 82)
ks = (ks - 1) .* ly.dilate + 1 ;

Error in vid_create_net>add_block (line 60)
info = vl_simplenn_display(net) ;

Error in vid_create_net>modified_alexnet (line 141)
net = add_block(net, opts, '2', 5, 5, 48, 256, strides(3), 0) ;

Error in vid_create_net (line 28)
net = modified_alexnet(struct(), opts) ;

Error in make_siameseFC (line 22)
branch = vid_create_net(...

Error in experiment>make_net (line 136)
net = make_siameseFC(opts);

Error in experiment (line 69)
net = make_net(opts);

Error in run_experiment (line 17)
experiment(imdb_video, opts);

[xpain1]

hi @bertinetto
i know why that happened.because my matconvnet is 1.0-21,which has a dilate in the conv layer.so in the vl_simplenn_display,the error occured. i change my natconvnet from 1.0-21 to 1.0-20,the code is succeed.