Hello
I followed the document to run SNPE
Example Tutorials -> Running Nets -> TensorFlow Inception v3
Running on Android using CPU Runtime
Although there are some errors in logcat, but works fine.
-------------------------------------------------------------------------------
Model String: N/A
SNPE v1.2.2.0
-------------------------------------------------------------------------------
Processing DNN Input: frame_43.bin
adb logcat -s adsprpc
--------- beginning of system
--------- beginning of main
01-01 00:07:57.924 3622 3622 E adsprpc : vendor/qcom/proprietary/adsprpc/src/fastrpc_apps_user.c:171::error: -5: 0 == (nErr = dlerr)
01-01 00:07:57.924 3622 3622 D adsprpc : vendor/qcom/proprietary/adsprpc/src/fastrpc_apps_user.c:174: dlopen error: snpe_dsp cannot find file or invalid file descriptor: fd=:-1
Running on Android using DSP Runtime
DSP runtime isn't available
The selected runtime is not available on this platform. Continue anyway to observe the failure at network creation time.
error_code=101; error_message=Invalid parameter in user config. Attempted to set a neural network configuration option DSP that is not supported on this platform.;
error_component=System Configuration; line_no=86; thread_id=-142763212
I checked adsprpcd:
ps | grep adsprpcd
media 718 1 8372 1256 poll_sched 7fa58c9454 S /system/bin/adsprpcd
logcat -s adsprpc
01-01 00:10:30.475 3749 3749 E adsprpc : vendor/qcom/proprietary/adsprpc/src/fastrpc_apps_user.c:171::error: -5: 0 == (nErr = dlerr)
01-01 00:10:30.475 3749 3749 D adsprpc : vendor/qcom/proprietary/adsprpc/src/fastrpc_apps_user.c:174: dlopen error: snpe_dsp cannot find file or invalid file descriptor: fd=:-1
Thanks
Asaku
APQ8096 DSP Runtime Unavailable
Posted: Fri, 2017-08-18 01:34
Posting this to bring the topic up.
I am having the same issue:
Does anyone know how to solve it?
SNPE version used in my installation is 1.8.0
UPD: the ADSP_LIBRARY_PATH is being set to a proper value in this way:
bool setAdspLibraryPath() {
return setenv("ADSP_LIBRARY_PATH",
}
, where /data/dsp contains the skel shared object (as advised in the documentation).