Malte J. Rasch

You need to set the flag CUDA_ARCHITECTURES correctly. For instance by: ``` make build_cuda flags='-DRPU_CUDA_ARCHITECTURES="86"' ``` Looks like the RTX A5000 has ``cuda_arch`` 86

That looks like that your CUDA library might not be installed properly? Or maybe try ``` make build_cuda flags='-DRPU_CUDA_ARCHITECTURES="86" -DCUDA_ARCHITECTURES="86"' ``` If that is not working make sure that it...

It cannot find your CUDA library installation: ``-- The CUDA compiler identification is unknown``. You need to make sure that nvcc is found and that CUDA library is properly installed...

11.7 is very new, it might have issues with that. In any case, your CUDA installation seems to have problems if it cannot find the version string in the cmake...

Hi @matifali , great that it worked out! Indeed, we are in the process of providing conda pre-builds for several CUDA versions. However, if you could contribute docker images or...

Hi @dingandrew, that is correct, the summed currents are the outputs. To "turn-off" all digital compute, you could set the `inp_res` and `out_res` to -1 to avoid any quantization. The...

We should add some kind of "AnalogOutputPreset" to the Presets to make this settings simpler.

@kaoutar55 The default is an assumption that we have digital outputs. Otherwise using activations etc in pytorch does not make much sense

You could simply add an analog pooling layer (with the expected noise sources etc) to your liking by adding a pytorch module and build a DNN using that custom pooling...

The `inp_bound` and `out_bound` should probably be a smaller number. Even for a purely analog circuit, you will have some output and input bounds as the currents will saturate at...