suriyadeepan / practical_seq2seq

A simple, minimal wrapper for tensorflow's seq2seq module, for experimenting with datasets rapidly
http://suriyadeepan.github.io/2016-12-31-practical-seq2seq/
GNU General Public License v3.0
570 stars 270 forks source link

tensorflow-gpu 1.12.0 uses 90% of CPU and 25%GPU. Not sure why #74

Open preethamsridhar opened 5 years ago

preethamsridhar commented 5 years ago

I am running 03-Twitter-chatbot.ipynb on my laptop which has NVIDIA 1080 8GB graphics card and 64GB i7 RAM processor. For the first 1000 iteration, it takes around half an hour. I believe it stops around 43000 iterations (from the pre-trained checkpoint tar,gz file i guessed).

I am not sure even though i am using tensorflow-gpu why is it using most of the CPU instead of GPU. Kindly let me know if you have the answer. FYI i am running on Windows 10 Professional.