safreita1 / resnet-emotion-recognition Goto Github PK

what is predict_value in dataset fer2013?
what is checkpoint error in code?
During handling of the above exception, another exception occurred:

NotFoundError Traceback (most recent call last)
NotFoundError: Key _CHECKPOINTABLE_OBJECT_GRAPH not found in checkpoint

During handling of the above exception, another exception occurred:

NotFoundError Traceback (most recent call last)
/usr/local/lib/python3.6/dist-packages/tensorflow/python/training/saver.py in restore(self, sess, save_path)
1300 # a helpful message (b/110263146)
1301 raise _wrap_restore_error_with_msg(
-> 1302 err, "a Variable name or other graph key that is missing")
1303
1304 # This is an object-based checkpoint. We'll print a warning and then do

NotFoundError: Restoring from checkpoint failed. This is most likely due to a Variable name or other graph key that is missing from the checkpoint. Please ensure that you have not altered the graph expected based on the checkpoint. Original error:
please avvise.

The current state of the art accuracy for FER2013 is at 75.2% to the best of my knowledge:

"Facial Expression Recognition using Convolutional Neural Networks: State of the Art":
https://pdfs.semanticscholar.org/4edc/7f27d4512b69be54abfc6b9876e5b00725ab.pdf

They use an ensemble of different CNN architecture such as ResNet, Inception & VGG.

The gap between 75% and the 94% you claim to get seems very large, are you sure you have evaluated your model correctly?