chiphuyen / stanford-tensorflow-tutorials

This repository contains code examples for the Stanford's course: TensorFlow for Deep Learning Research.
http://cs20.stanford.edu
MIT License
10.32k stars 4.32k forks source link

Training Data Disappears #150

Open limburgercheese opened 4 years ago

limburgercheese commented 4 years ago

I am finding that as the bot trains the older data outside of the last 10ish steps disappears. Is this normal? I would imagine that the bot would need large volumes of data in order to function but it always stays around 2.5 GB which doesn't sound right. Please advise.