I am new to caffe, and after running the example successfully, I try to use my own data. However, trying to either write my data to the lmdb data format or to try to use the solver directly, in both cases I get an error:
E0201 14:26:00.450629 13235 io.cpp:80] Could not open or find file ~/Documents/ChessgameCNN/input/train/731_1.bmp 731
The path is right, but it’s strange that the label 731is part of this error message. This means that he reads it as part of the path, and not as a label. The text file is as follows:
~/Documents/ChessgameCNN/input/train/731_1.bmp 731
Is it because the tags are too high? Or maybe because labels do not start at 0? I searched for this error, and all I found was examples with a relatively small number of labels, about ~ 1-5, but I have about 4096 classes in which I do not always have examples in the training data. Perhaps this is also a problem (of course, for training, at least, but I did not expect this to give me the actual error message). Typically, the label is not part of this error message. To create the lmdb file, I use the create_imagenet.shcaffe examples. For the solution, I use:
~/caffe/build/tools/caffe train --solver ~/Documents/ChessgameCNN/caffe_models/caffe_model_1/solver_1.prototxt 2>&1 | tee ~/Documents/ChessgameCNN/caffe_models/caffe_model_1/model_1_train.log
I also tried different types of image data: PNG, JPEG and BMP. So this is not a criminal either. If this is really because of my choice of labels, what would be a viable workaround for this problem?
Many thanks for your help!