search

thomas-koehler / SupER

Super-Resolution Erlangen (SupER): Benchmarking Super-Resolution Algorithms on Real Data
MIT License
79 stars 31 forks source link

Cannot create 'trees_x2_ibp_2048_n16.mat' #4

Closed makslevental closed 5 years ago

makslevental commented 5 years ago 
Error using save
Cannot create 'trees_x2_ibp_2048_n16.mat' because
'/home/maksim/dev_projects/SupER/matlab/algorithms/SRAlgorithms/NBSRF/private_folder/../models'
does not exist.

Error in nbsrf_training

Error in superresolve_nbsrf (line 23)
    setup = nbsrf_training(setup);

Error in callSRMethod (line 7)
    srImage = feval(funName, slidingWindow, magFactor);

Error in runSRAlgorithm (line 30)
    srImage.(srMethod.name) = callSRMethod(slidingWindow, magFactor,
    srMethod);

Error in processDatasets (line 70)
                                    [srImages, timeME, timeSR] =
                                    runSRAlgorithm(imageDataset.(['Bin',
                                    num2str(binningFactors(binningFactorIdx))]),
                                    imageDataset.groundTruth,
                                    binningFactors(binningFactorIdx),
                                    startReferenceFrame,
                                    numberOfFrames(numberOfFramesIdx,
                                    binningFactorIdx), slidingIdx,
                                    sr_method);

Error in evaluateBaseline (line 21)
processDatasets(inputDir, datasetConfig, resultDir, binningFactors,
numberOfFrames, startReferenceFrame, numberOfFrames_val, sliding_val,
sr_method, binning_val, scenes_val, compressions(compress_val));
thomas-koehler commented 5 years ago

Hi,

This error appears because the folder containing the NBSRF models were missing. I excluded the models due to their file size. If you run NBSR for the first time, it will re-train and save the models in that folder. I added the required models folder without the model files to solve this issue.

Best,

Thomas