I ran tensorflow object detection api tutorial notebook and got accuracy show in image using 'ssd_mobilenet_v1_coco_2017_11_17' model which is different from the tutorial i am following(lower to be precise).
My question is does a model accuracy vary from system to system.
I'm running this on window 10 pc with amd processsor(CPU only traning).
There is a lot of things that can influence your inference. The model of your GPU or CPU is not one of them. But many things can be different because you are using your CPU, like memory usage, batchsize and other things.
Tensorflow and a lot of libraries used in this object detection api tutorial has changed from 2017 (date of the video), so for a 1:1 detection you need to install the same 2017 version of the tensorflow and all of it's requirements.
