Article: Yoo S y J C y, Yun K, Choi J y. Action-decision Networks for Visual Tracking with deep reinforcement learning[c]//cvpr2 017
Error 1:undefined function or variable ' Test_demo '.
Solution: This is very simple, directly to change the Test_demo into Adnet_test can
Error 2:Invalid mex-file xxxxxxxxxxxxx\utils\croprectanglesmex.mexw64': The specified module could not be found.
Workaround: This is because the CROPRECTANGLESMEX.MEXW64 in the Utils\croprectanglesmex directory needs to be recompiled, read the Readme in the directory before compiling it and find out that this code is from here https:// Github.com/aosokin/croprectanglesmex
However, the source code is compiled under the Linux system, and the author of Adnet gives the compiler and linker under Windows, so if it is a Windows system, you need to run the BUILD_CROPRECTANGLESMEX_ON_WINDOWS.M file
When you run the file, you may find a hint similar to the errorNVCC fatal:cannot find compiler ' cl.exe ' in PATH
Where Cl.exe is the compiler for Visual studio, so you need to add Cl.exe directory to the environment variable, my cl.exe directory is: D:\Program Files (x86) \microsoft Visual Studio 11.0\vc\bin
It is also important to note that the Cuda version given by BUILD_CROPRECTANGLESMEX_ON_WINDOWS.M is 7.5 and needs to be adjusted according to its own version number. For example, I changed the code for LINE10: cudaroot = ' C:\Program files\nvidia GPU Computing toolkit\cuda\v8.0 ';
If the BUILD_CROPRECTANGLESMEX_ON_WINDOWS.M runs successfully, do not forget to paste croprectanglesmex.mexw64 from the Utils\croprectanglesmex directory into the Utils\ Directory, the existing CROPRECTANGLESMEX.MEXW64 file has been replaced.
Adnet Code Issues