i was playing around with the emotion data set on kaggle using tensorflow. depending on the seed, i was getting between 58 and 60% accuracy on the hold out tes

i was playing around with the emotion data set on kaggle using tensorflow. depending on the seed, i was getting between 58 and 60% accuracy on the hold out test set (what you submit against). i thought i came up with a good set of hyper parameters using aws gpu instances (python 2.7). i wanted to visualize some of the outputs so I copied the code to my machine and ran under python 3.5 (windows) and only got 57% accuracy. these swings in accuracy are huge