Issue freezing graph
mdlockyer opened this issue · 1 comments
I've been testing out this implementation and it's quite good! I do have a slight issue though.
I'm trying to freeze the graph into a pb file, however I cant seem to ascertain the proper output node name. From my understanding, the output should be conv6
, however when I try to convert the vars to constants the output name conv6
isn't found in the graph. Any ideas?
Hello @MichaelDylan77,
In the inference script you will see that the output of the network before passing the information to the decoder is :
raw_output_up = tf.argmax(raw_output_up, axis=3)
Converting the ckpt to tf.events and visualizing it to tensorboard you will see that the output of the network is : ArgMax layer without scope name.
Don't forget to call the decoder : pred = decode_labels(raw_output_up, img_shape, num_classes)
after inference using the .pb file.