- Use dvt data to train the model
- Fall back to vskin-speaker-legacy:
- if all samples not collected
- if vskin-speaker-model temp < 20degC
- if vskin-speaker-model temp >= 57degC
- Add timeResolution to avoid duplicate readings issue.
Bug: 343157242
Test: Push config and tflite model to verify lshal and logcat
Change-Id: I2140a92ae3b19825ec9542e3ec65a365bec2e66f