How to predict(execute object detection) with trained model ?

[JudeLee19]

Hi, I'm new in torch. I followed your readme.md and having a success training vgg model with ImageNet data. But when I call evaluation_demo function for test images with create snapshots such as imgnet_018000.t7, imgnet_019000.t7 ... I got below error messages .

main.lua:72: bad argument #1 to 'copy' (sizes do not match at /tmp/luarocks_cutorch-scm-1-2930/cutorch/lib/THC/generic/THCTensorCopy.cu:10)
stack traceback:
[C]: in function 'copy'
main.lua:72: in function 'load_model'
main.lua:164: in function 'evaluation_demo'
main.lua:197: in main chunk
[C]: in function 'dofile'
.../dev/torch/install/lib/luarocks/rocks/trepl/scm-1/bin/th:145: in main chunk
[C]: in ?

And the part that error is raised is weights:copy(stored.weights) in main.lua file.
local stored = load_obj(network_filename)
training_stats = stored.stats
weights:copy(stored.weights)

Thank u in advance.

[JudeLee19]

I made a mistake. In training step I used vcc_large but when test step I used vcc_small .. that's why each size is different.