About Training Details.

[NeverMoreLCH]

trafficstars

Hi,

Your work is really great! I have some questions when I train the cfnet-conv2 network. In your code 'run_experiment_cfnet_conv2.m', the epoch is set to 100. But in your pretrained networks, the pretrained networks for cfnet-conv2 are called 'cfnet-conv2_e80.mat' and 'cfnet-conv2_gray_e40.mat'. I want to know how many epochs do you set when training the cfnet-conv2 network (color and gray).

Thank you very much! Hope your reply!

[bertinetto]

Hi, This is the protocol we adopted, which we used to avoid overfitting to the test set(s).

• We train CFNet/SiamFC for 100 epochs
• We measure its tracking performance (on validation set) every 5 epochs (or 10, I can't remember).
• To report the results in plots (with error bars), we consider the variance of the last 10 (evaluated) epochs.
• To report final results and compare against the state-of-the-art, we pick the epoch of the network that maximizes the performance on the validation set and then run it on test set (e.g OTB/LaSoT/GOT-10k...). That epoch is reported in the mat file id (e.g. e40 or e80).

Hope this helps!