Snpe 1.68 failed to use aip
Join Date: 5 Feb 23
Posts: 3
Posted: Sun, 2023-02-05 22:45
Hello,
I tried to use the snpe 1.68 version of Snapdragon 865 to call the cpu, gpu, and dsp to run the yolo model, and successfully output the expected results.
However, I try to use the aip running yolo model in the snpe 1.68 version in the Snapdragon 865, and got the following error
model_name: yolo_quantized_htp.dlc
img_path path: /data/local/tmp/infer_resize/
infer_type 4
runtime 4
4 AIP_FIXEDB_TF
runtime_t : aip_fixed8_tf
SNPE model init, using runtime aip_fixed8_tf
PlatformOptions (unsignedPD:ON) set successful config option is valid
npu_get_property status: 0
npu_get_property status: 0
FW CAPS [0] = 0x2007
FW CAPS [1] = 0x80000005
FW CAPS [2] = 0x0
FW CAPS [3] = 0x0
FW CAPS [4] = 0x0
FW CAPS [5] = 0x0
FW CAPS [6] = 0x0
FW CAPS [7] = 0x0
npu_get_property status: 0
NPU User Driver: npu_read_info 0
npu_get_property status: 0
npu_get_property status: 0
FW CAPS [0] = 0x2007
FW CAPS [1] = 0x80000005
FW CAPS [2] = 0x0
FW CAPS [3] = 0x0
FW CAPS [4] = 0x0
FW CAPS [5] = 0x0
FW CAPS [6] = 0x0
FW CAPS [7] = 0x0
npu_get_property status: 0
build engine error : error_code=500; error_message=Target runtime is not available. error_code=500; error_message=Target runtime is not available. No viable runtimes available.; error_component=Host Runtime; line_no=650; thread_id=548097474256; error_component=Host Runtime; line_no=342; thread_id=548097474256
Do you have any idea about this problem ?
Dear developer,
What's the conversion commands for your quantization model flow? You need to add option of --use_aip if target runtime is dsp.
BR.
Wei
Hello,
My conversion command is snpe-dlc-quantize --input_dlc yolov3.dlc --input_list image_file_list.txt --output_dlc yolov3_quantize --enable_hta. but I run the official API c++ code and set zdl::DlSystem::Runtime_t::AIP_FIXED8_TF.
can you help me?
thank you