dBeker / Faster-RCNN-TensorFlow-Python3

Tensorflow Faster R-CNN for Windows/Linux and Python 3 (3.5/3.6/3.7)
MIT License
612 stars 329 forks source link

I want to compiled cython modules with sm75 architecture (RTX2080TI) #81

Closed gwspotex closed 5 years ago

gwspotex commented 5 years ago

Can you teach me how to change architecture? image

morpheusthewhite commented 5 years ago

You just need to follow instructions in the numbered list (install tensorflow-gpu)

gwspotex commented 5 years ago

sorry, but where is the numbered list? instructions?

gwspotex commented 5 years ago

there are question&answer: https://github.com/endernewton/tf-faster-rcnn/issues/5 it seems the dismatch architecture makes the result get some loss,i don't how complie the sm61(gtx 1080ti)->sm75(rtx2080ti)

morpheusthewhite commented 5 years ago

In the README

morpheusthewhite commented 5 years ago

there are question&answer: endernewton/tf-faster-rcnn#5 it seems the dismatch architecture makes the result get some loss,i don't how complie the sm61(gtx 1080ti)->sm75(rtx2080ti)

That's another project, contact its mantainer

gwspotex commented 5 years ago

ok :),because i see the readme say there sm61 architecture and i just want using sm75 to change image

morpheusthewhite commented 5 years ago

You are looking at some commit while you should read directly from the README