search

localminimum / QANet

A Tensorflow implementation of QANet for machine reading comprehension
MIT License
983 stars 310 forks source link

How to train in Multi GPU #31

Open deepakkumar1984 opened 6 years ago

deepakkumar1984 commented 6 years ago

I see that tensorflow detected 2 GPU's but the training is only happening in 1 GPU. Please advise?

odellus commented 6 years ago

It's not automatic. My 2 cents: look at this and figure out how you want to do it. https://www.tensorflow.org/programmers_guide/using_gpu#using_multiple_gpus

localminimum commented 6 years ago

Hi @deepakkumar1984 , multi-gpu training is not on my priority list, as SQuAD models usually train in a couple of hours with affordable gpus (gtx1070 etc). But I will definitely add it to the todo list for future work. Even better, a pull request and contributions are always welcome.