Proposals t7 files
minouminou opened this issue · 1 comments
Hello how to generate proposals t7 files from deepmask output on my data? I use code in issue #16. When I try to train on 1 GPU I see error:
/home/alex/torch/install/bin/luajit: ./modules/BBoxRegressionCriterion.lua:27: bad argument #1 to 'copy' (sizes do not match at /tmp/luarocks_cutorch-scm-1-4541/cutorch/lib/THC/THCTensorCopy.cu:31) stack traceback: [C]: in function 'copy' ./modules/BBoxRegressionCriterion.lua:27: in function 'updateOutput' ...lex/torch/install/share/lua/5.1/nn/ParallelCriterion.lua:23: in function 'forward' ./engines/fboptimengine.lua:61: in function 'train' train.lua:363: in main chunk [C]: in function 'dofile' ...alex/torch/install/lib/luarocks/rocks/trepl/scm-1/bin/th:150: in main chunk [C]: at 0x00405d50
If I fill all numbers in all bboxes with same number (ex. [10 10 10 10]) training loads all CPUs (100%), but not GPU (0%). Looks strange
Thanks for any hints.
Was need to change row opt.num_classes = opt.dataset == 'pascal' and 21 or 81 to opt.num_classes = 2
to opt.num_classes=<number of classes+1>