Bruno Barbosa

I did face this issue for _val_loss_ when disabled the random flag on the validation data generator

Hi @fffupeng! I use an additional script to do that. ```""" Tensorflow graph freezer Converts Tensorflow trained models in .pb Code adapted from: https://gist.github.com/morgangiraud/249505f540a5e53a48b0c1a869d370bf#file-medium-tffreeze-1-py """ import os, argparse os.environ['TF_CPP_MIN_LOG_LEVEL'] =...

Hi @thenathan30 ! Can you show me a snapshot of your directory tree where your model is stored?

Hi @MidsizeMango! You should visualize your network on the TensorBoard and check which layer you want to export. In this case (classification problem) i need the output of the softmax...

Hi @poojashah89! Take a look on this [link](https://gist.github.com/morgangiraud/249505f540a5e53a48b0c1a869d370bf#gistcomment-2195007) In my case, I have a the softmax function is used on the third fully_connected layer. So, for me `output_node_names = "FullyConnected_2/Softmax"`

Hi @kucinghitam13! I'm glad you made it. Did you apply any preprocessing methods in your training? If you did, you must replicate them as well.

As far as I know, TFLearn can't run frozen graphs. Currently, I use frozen graphs on: - OpenCV DNN Module - TensorFlow Mobile - TensorFlow (general) Otherwise, I use the...

Hi @fly2015! As I mentioned on https://github.com/tflearn/tflearn/issues/964#issuecomment-350717215, try to do `del tf.get_collection_ref(tf.GraphKeys.TRAIN_OPS)[:]` before saving model.

You are welcome @fly2015 From your code I see that you forget to include the word `del` before `tf.get_collection_ref(tf.GraphKeys.TRAIN_OPS)[:]` Can you try it again please?

Hi @fly2015! I'm glad you made it. At the moment I'm using frozen models on an Android app. However, I never tried it on Java. Best regards.