to-be-snail

> Can you try `FullPrecLearner` (no compression)? It seems that after pruning,there are some differences between training model and evaluating model.

> Okay, so can you post the training and evaluation models you generated? [-10000.zip](https://github.com/Tencent/PocketFlow/files/2714023/-10000.zip) INFO:tensorflow:model saved to ./models/models_eval_mobilenet_v1_at_cifar10/models/-10000 INFO:tensorflow:Restoring parameters from ./models/models_eval_mobilenet_v1_at_cifar10/models\-10000 INFO:tensorflow:model restored from ./models/models_eval_mobilenet_v1_at_cifar10/models\-10000 when pruning,the trained model...

Hello，can I ask a few questions about built caffe-window with VS2017 and cuda10.0 @wuqingshan2010

https://blog.csdn.net/mao_hui_fei/article/details/80326464

yeah,the prob and prediction is different. In addition,when I convert a unet keras model to caffe model ,also get the wrong result,so I trying to observe the intermediate layer output...

> What do you mean as different result? Wrong class prediction? yeah,the prob and prediction is different. In addition,when I convert a unet keras model to caffe model ,also get...

> The small difference is present due to layers different implementation in tf and caffe, as noted here #3. > But overall Keras VGG16 model conversion works well for me...

> using I'm using keras2.2.4,the backend is tensorflow 1.9.0, caffe is 1.0 and built on python3.6

> It seems that the output of convolution layers of keras and caffe has different output even when the input the and weights are all the same. It may be...

> I think pooling layer has the difference between caffe and tf I met the same question.Did you solve the problem?