我正在使用anaconda在Ubuntu 14.04上运行caffe。
当我运行examples/mnist/train_lenet.sh
./build/tools/caffe: /home/kli/anaconda/lib/libtiff.so.5: no version information available (required by /home/kli/local/lib/libopencv_highgui.so.2.4) ./build/tools/caffe: /home/kli/anaconda/lib/liblzma.so.5: no version information available (required by /usr/lib/x86_64-linux-gnu/libunwind.so.8) I0303 22:05:06.373118 11808 caffe.cpp:185] Using GPUs 0 I0303 22:05:08.649866 11808 caffe.cpp:190] GPU 0: GeForce GTX TITAN X
我不知道出了什么问题。 这似乎是一些警告,因为培训过程没问题。
I0303 22:11:11.745873 11855 solver.cpp:228] Iteration 100, loss = 0.227769 I0303 22:11:11.745945 11855 solver.cpp:244] Train net output #0: loss = 0.227769 (* 1 = 0.227769 loss) I0303 22:11:11.745960 11855 sgd_solver.cpp:106] Iteration 100, lr = 0.00992565 I0303 22:11:11.943850 11855 solver.cpp:228] Iteration 200, loss = 0.155188 I0303 22:11:11.943912 11855 solver.cpp:244] Train net output #0: loss = 0.155188 (* 1 = 0.155188 loss) I0303 22:11:11.943928 11855 sgd_solver.cpp:106] Iteration 200, lr = 0.00985258 I0303 22:11:12.132534 11855 solver.cpp:228] Iteration 300, loss = 0.160685 I0303 22:11:12.132566 11855 solver.cpp:244] Train net output #0: loss = 0.160685 (* 1 = 0.160685 loss) I0303 22:11:12.132580 11855 sgd_solver.cpp:106] Iteration 300, lr = 0.00978075 I0303 22:11:12.320441 11855 solver.cpp:228] Iteration 400, loss = 0.0642169 I0303 22:11:12.320477 11855 solver.cpp:244] Train net output #0: loss = 0.064217 (* 1 = 0.064217 loss) I0303 22:11:12.320490 11855 sgd_solver.cpp:106] Iteration 400, lr = 0.00971013 I0303 22:11:12.528359 11855 solver.cpp:337] Iteration 500, Testing net (#0) I0303 22:11:12.635912 11855 solver.cpp:404] Test net output #0: accuracy = 0.972 I0303 22:11:12.635975 11855 solver.cpp:404] Test net output #1: loss = 0.0880498 (* 1 = 0.0880498 loss) I0303 22:11:12.636888 11855 solver.cpp:228] Iteration 500, loss = 0.121124 I0303 22:11:12.636927 11855 solver.cpp:244] Train net output #0: loss = 0.121124 (* 1 = 0.121124 loss) I0303 22:11:12.636947 11855 sgd_solver.cpp:106] Iteration 500, lr = 0.00964069 I0303 22:11:12.827488 11855 solver.cpp:228] Iteration 600, loss = 0.0760444 I0303 22:11:12.827522 11855 solver.cpp:244] Train net output #0: loss = 0.0760446 (* 1 = 0.0760446 loss)