2018-12-30 20:55:21 -0600 | edited question | DNN module: switch Torch model between train() and evaluation() modes DNN module: set Torch model to evaluation() mode Torch models can run in two different modes: train() and evaluation(). |
2018-12-21 00:59:10 -0600 | commented answer | DNN module: switch Torch model between train() and evaluation() modes @dkurt Thanks for the pull request! |
2018-12-21 00:56:05 -0600 | received badge | ● Enthusiast |
2018-12-20 23:52:08 -0600 | commented answer | DNN module: switch Torch model between train() and evaluation() modes @dkurt If I could, I would give you 1000000 upvotes. Thank you so much for your kind feedback! |
2018-12-20 23:21:17 -0600 | commented answer | DNN module: switch Torch model between train() and evaluation() modes @dkurt I know that some layers, such as dropout or batch normalization are not deterministic during the training phase. |
2018-12-20 21:04:43 -0600 | marked best answer | DNN module: switch Torch model between train() and evaluation() modes Torch models can run in two different modes: train() and evaluation(). Results of some layers such as batch normalization will be affected by the modes. Currently it seems (i.e., when I tested) that OpenCV's DNN module loads and runs torch models in train() mode. Is there a way to use torch models in evaluation() mode using OpenCV? (ADDED) Below is the lua code for running the model with two different modes. |
2018-12-20 21:04:43 -0600 | received badge | ● Scholar (source) |
2018-12-20 21:04:37 -0600 | commented answer | DNN module: switch Torch model between train() and evaluation() modes Thanks for your answer! I also got same result by using OpenCV for Python. May I ask one more question? Then is there a |
2018-12-20 20:56:02 -0600 | commented answer | DNN module: switch Torch model between train() and evaluation() modes Thanks for your answer. I also got same result by using OpenCV for Python. But my code is on Android and the result from |
2018-12-20 03:41:35 -0600 | edited question | DNN module: switch Torch model between train() and evaluation() modes DNN module: set Torch model to evaluation() mode Torch models can run in two different modes: train() and evaluation(). |
2018-12-20 03:39:49 -0600 | commented question | DNN module: switch Torch model between train() and evaluation() modes @dkurt The model is on https://github.com/chuanli11/MGANs/blob/master/model/Picasso.t7 . (ADDED) I added the my lua code |
2018-12-20 03:35:46 -0600 | commented question | DNN module: switch Torch model between train() and evaluation() modes @dkurt The model is on https://github.com/chuanli11/MGANs/blob/master/model/Picasso.t7 . |
2018-12-20 03:31:26 -0600 | commented question | DNN module: switch Torch model between train() and evaluation() modes @dkurt Oh, I used net:evaluate(), not net:evaluation(), in torch. Sorry for the typo. I am sure the inputs were exactly |
2018-12-20 03:27:40 -0600 | commented question | DNN module: switch Torch model between train() and evaluation() modes @dkurt Oh, I used net:evaluate(), not net:evaluation(), in torch. Sorry for the typo. I am sure the inputs were exactly |
2018-12-20 03:19:16 -0600 | commented question | DNN module: switch Torch model between train() and evaluation() modes @dkurt I have a deconvolution network. I ran it on my desktop with same input, once with model:train() and the other tim |
2018-12-20 02:59:01 -0600 | commented question | DNN module: switch Torch model between train() and evaluation() modes @dkurt Would you explain to me a bit in detail? I tried calling model:evaluation() and then using torch.save() to save t |
2018-12-20 02:42:59 -0600 | asked a question | DNN module: switch Torch model between train() and evaluation() modes DNN module: set Torch model to evaluation() mode Torch models can run in two different modes: train() and evaluation(). |
2018-11-08 02:31:40 -0600 | commented question | Is there implicit limit for DNN input size of OpenCV Android? @berak Testing on a PC rather than Android seems a good idea. Thanks! |
2018-11-07 20:19:55 -0600 | commented question | Is there implicit limit for DNN input size of OpenCV Android? @berak I added the prototxt content to the post. |
2018-11-07 20:19:27 -0600 | received badge | ● Editor (source) |
2018-11-07 20:19:27 -0600 | edited question | Is there implicit limit for DNN input size of OpenCV Android? Is there implicit limit for DNN input size of OpenCV Android? I trained a Caffe model which does NOT specify input sizes |
2018-11-07 03:15:40 -0600 | commented question | Is there implicit limit for DNN input size of OpenCV Android? @berak It's OpenCV 3.4.2 for Android. The network is a caffe network without specific input size. |
2018-11-07 03:01:49 -0600 | received badge | ● Organizer (source) |
2018-11-07 03:01:43 -0600 | asked a question | Is there implicit limit for DNN input size of OpenCV Android? Is there implicit limit for DNN input size of OpenCV Android? I trained a Caffe model which does NOT specify input sizes |