Dear Comunnity,
Hope you had a great easter. I would like to play around speech recognition/ speech processing with Nengo (esp Nengo-FPGA). Do we have any example for that?
Sincerely
Murat
Hi @muratisik,
We don’t have many examples working with speech recognition / processing, but I did manage to find one example here (it’s with the Loihi backend, but it should work in “standard” Nengo as well).
We definitely do not have any examples of speech recognition / processing with NengoFPGA because that platform is much more restrictive than standard Nengo. With NengoFPGA, only one Nengo ensemble can be implemented per FPGA (you can connect multiple FPGAs to your PC if you want to have multiple ensembles implemented on the FPGA - but still, one ensemble per FPGA), and these speech recognition / processing networks are typically multi-layered.
Hi Dr. @xchoo,
Thank you so much for your reply. I will look at this example. I am not sure do I have an access to Loihi. It would be nice to see any example on FPGA. This tool (https://modelzoo.co/model/espresso) is looking quite popular for speech recognition.
Hi @muratisik,
If you use our NengoLoihi package, you do not need to have a physical Loihi board connected. If you set the target="sim" when you create the nengo_loihi.Simulator object, it will run the Nengo model on your PC in a Loihi hardware “emulator” (see the documentation here). As for the FPGA example, unfortunately, we don’t currently have the resources to explore making one, so we have to leave it up to the user to do their own exploration with the Nengo software. 