Hi. I successfuly compiled two dependencies on Ubuntu 16.04 with CUDA 10.0 and MATLAB R2019b. There seems to be no issue running the MatConvnet demo. But I got this error when I am trying to run run_experiments.m
Error using vl_argparse (line 52)
The name of the parameter number 1 is not a string.
Error in vl_nnrelu (line 41)
opts = vl_argparse(opts, varargin, 'nonrecursive') ;
Error in vl_simplenn (line 336)
res(i+1).x = vl_nnrelu(res(i).x,[],leak{:}) ;
Error in get_dcnn_features (line 111)
res = vl_simplenn(net, im_resized, [], res, ...
Error in os_train>get_batch_results (line 319)
code_ = get_dcnn_features(encoder.net, ...
Error in os_train>encoder_extract_for_segments (line 269)
parfor (b = 1:numel(batches), numWorkers)
Error in os_train>encoder_train_from_segments (line 393)
code = encoder_extract_for_segments(encoder, imdb, segmentIds) ;
Error in os_train (line 27)
encoder = encoder_train_from_segments(...
Error in run_experiments (line 109)
os_train(...
Is it possible to fix it? Thank you very much for your time!
Sorry. I figured out should be vlfeat's issue.
I used vlfeat v0.9.20 to compile. Replacing vlfeat/toolbox/misc/vl_argparse.m with a newer version solved this issue.
Hi. I successfuly compiled two dependencies on Ubuntu 16.04 with CUDA 10.0 and MATLAB R2019b. There seems to be no issue running the MatConvnet demo. But I got this error when I am trying to run
run_experiments.m
Is it possible to fix it? Thank you very much for your time!