Closed PeterPan1990 closed 9 years ago
Here's a snapshot at 20K iterations:
https://s3.amazonaws.com/jgoode/oxford102_VGG_S_iter_20000.caffemodel
which gives
1108 15:07:57.676288 1699 solver.cpp:266] Iteration 20000, Testing net (#0)
I1108 15:08:55.986631 1699 solver.cpp:315] Test net output #0: accuracy = 0.827742
I1108 15:08:56.456195 1699 solver.cpp:189] Iteration 20000, loss = 0.00186366
I1108 15:08:56.456280 1699 solver.cpp:464] Iteration 20000, lr = 0.0001
Hi, Thanks for sharing the experiments info, I was wondering that if could give me a link to pretrained-weight for VGG_S finetune part?