As it is well known, TensorFlow needs at least compute capability higher than 3.5 to use gpu accelerations.
in hardware requirements mentioned in tensorflow.org, it was stated that :
For GPUs with unsupported CUDA® architectures, or to avoid JIT
compilation from PTX, or to use different versions of the NVIDIA®
libraries, see the Linux build from source guide.
I have seen several cases which build tensorflow for graphic card of 3.0 compute capability(cc). I am eager to know does this method work for lower cc(=2.2) like Nvidia Quadro 1000M?
Thanks in advance