I'm starting to play around with python neural networks, mainly for object recognition like this one: https://github.com/huangshiyu13/RPNplus. I plan to train more networks with more kinds of objects.

So far, this network seems to take a long time (~42hrs and counting) for training or maybe my CPU (16GB RAM, Intel i5) its not enough.

A friend of mine is selling a GPU NVidia GeForce 760 with 4GB RAM and I wonder if it is a good opportunity for me or if really a GPU is not a real advantage for NN, as I read in this other post.

What is your opinion? Is a GPU a good investment or not really? I've seen other GPU like ASUS with 2 or 4 GB for around 200 US$, does this NVidia make a big difference or any other common GPU will do the same work? (thinking in how much I should pay for this).


Buying an NVIDIA GPU for AI is recommended as most libraries require the CUDA toolkit to accelerate computation. Some frameworks like TensorFlow only support NVIDIA GPUs for enhanced performance so its the only card that is worth investing in at the moment

|improve this answer|||||

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.