Hi all,
I’ve compiled Kaldi source code on my JETSON NANO DEVELOPER KIT , model P3450.
I didn’t use NVIDIA’s docker since I’m using my own kaldi repository.
I’ve tried to perform recognition of one audio file with cuda decoder but the process was killed by the OS ( Ubuntu 18.04 ).
I would like to ask about the parameters I should pass to the binary to make it work on NANO device.
I’m using a TDNN-F model of size 70mb and a language model of size 1mb.
Thanks,