search

jimgoo / caffe-oxford102

Caffe CNNs for the Oxford 102 flower dataset
127 stars 54 forks source link

can you tell me ? #8

Open xiaozhuang1010 opened 7 years ago

xiaozhuang1010 commented 7 years ago

I use the same way as you and obtained a good performance,but when I use a single picture to test the model it only obtained 0.0098 accuracy.....please tell me why？

jimgoo commented 7 years ago

I'm not sure I understand. Are you saying you fed each image into the model individually, then recorded the class prediction, then divided the number of correct predictions by the number of images you fed in and got 0.0098?

xiaozhuang1010 commented 7 years ago

I means enter a single image to the trained model(oxford102_VGG_S_iter_20000.caffemodel) and obtained equal probability of classification.just like: sweet pea:0.0098 english marigold:0.0098 pink primross：0.0098....and so on.

frank-wen commented 7 years ago

hi, I got exactly the same problem, have you figured it out?

frank-wen commented 7 years ago

I solved it. Just need to edit the deploy file properly