This Android application is meant to accompany the deep learning server to create a full deep learning cloud offload demonstration.
See the documentation for the server to setup the server-side of this demonstration.
It records audio then sends it to the server for inference, displaying the label returned by the server.
It also offers the user a chance to suggest an alternate label for the recorded audio, in order to demonstrate the re-training functionality of the deep learning server.
-
Edit the string resource
dl_server_uri
to point to the address of the deep learning server, e.g.dns:///my.address:55221
. -
Start the deep learning server.
-
Start the application and hit the record button!
- The "virtual microphone uses host audio input" option may need to be checked if running in an emulator. See the Android emulator settings for more information.