Segmentation fault when running example script or custom models on usb coral

[Skillnoob]

Description

After running the example script or ultralytics example i get a segmentation fault.
I followed the installation instructions for the USB coral and did so multiple times making sure i had nothing wrong.
It worked previously (few months ago) but does not anymore.
Video of it working correctly:

2023-09-16_13-27-48_Trim.mp4

Some information about the setup:
Os: Rasberry Pi Os Bullseye 64bit / Bookworm 64bit
Device: Rasberry Pi 4B 2GB / Raspberry Pi 5 8GB
tflite_runtime version: 2.10 until latest have been tested
Python: 3.9.2 (Pi 4B) / 3.11.2 (Pi 5) (3.7.16 and 3.9.18 have also been tested through conda)

Ways to reproduce the issue:
Run the detection example script from here.

Or follow these steps for a custom model:

$ pip install ultralytics

$ yolo export model=yolov8n.pt format=edgetpu

$ yolo detect val model=yolov8n_saved_model/yolov8n_full_integer_quant_edgetpu.tflite data=coco128.yaml batch=1

Click to expand!

Issue Type

Bug

Operating System

Linux

Coral Device

USB Accelerator

Other Devices

Rapsberry Pi 4

Programming Language

Python 3.9, Other

Relevant Log Output

When run through gdb:

Starting program: /usr/bin/python3 detect_image.py \
  --model models/ssd_mobilenet_v2_coco_quant_postprocess_edgetpu.tflite \
  --labels models/coco_labels.txt \
  --input images/grace_hopper.bmp \
  --output images/grace_hopper_processed.bmp
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/aarch64-linux-gnu/libthread_db.so.1".
[New Thread 0x7ff547e1c0 (LWP 4159)]
[New Thread 0x7ff4c7d1c0 (LWP 4160)]
[New Thread 0x7ff047c1c0 (LWP 4161)]
[New Thread 0x7fec6ee1c0 (LWP 4162)]
[Thread 0x7fec6ee1c0 (LWP 4162) exited]
[New Thread 0x7fec6ee1c0 (LWP 4163)]
[Thread 0x7fec6ee1c0 (LWP 4163) exited]
[New Thread 0x7fec6ee1c0 (LWP 4164)]
[New Thread 0x7febeed1c0 (LWP 4165)]
[New Thread 0x7feb6ec1c0 (LWP 4166)]
[Thread 0x7feb6ec1c0 (LWP 4166) exited]
[Thread 0x7febeed1c0 (LWP 4165) exited]
[New Thread 0x7febeed1c0 (LWP 4167)]
[New Thread 0x7feb6ec1c0 (LWP 4168)]
[Thread 0x7feb6ec1c0 (LWP 4168) exited]
[Thread 0x7febeed1c0 (LWP 4167) exited]
[New Thread 0x7febeed1c0 (LWP 4185)]
[New Thread 0x7feb6ec1c0 (LWP 4186)]
[New Thread 0x7feaeeb1c0 (LWP 4187)]

Thread 1 "python3" received signal SIGSEGV, Segmentation fault.
0x0000007fecb89350 in tflite::Subgraph::OpInit(TfLiteRegistration const&, char const*, unsigned long) () from /home/pi/.local/lib/python3.9/site-packages/tflite_runtime/_pywrap_tensorflow_interpreter_wrapper.so



Using backtrace from gdb:

Thread 1 "python3" received signal SIGSEGV, Segmentation fault.
0x0000007fecb89350 in tflite::Subgraph::OpInit(TfLiteRegistration const&, char const*, unsigned long) () from /home/pi/.local/lib/python3.9/site-packages/tflite_runtime/_pywrap_tensorflow_interpreter_wrapper.so
(gdb) backtrace
#0  0x0000007fecb89350 in tflite::Subgraph::OpInit(TfLiteRegistration const&, char const*, unsigned long) () from /home/pi/.local/lib/python3.9/site-packages/tflite_runtime/_pywrap_tensorflow_interpreter_wrapper.so
#1  0x0000007fecb8c4dc in tflite::Subgraph::AddNodeWithParameters(std::vector<int, std::allocator<int> > const&, std::vector<int, std::allocator<int> > const&, std::vector<int, std::allocator<int> > const&, char const*, unsigned long, void*, TfLiteRegistration const*, int*) () from /home/pi/.local/lib/python3.9/site-packages/tflite_runtime/_pywrap_tensorflow_interpreter_wrapper.so
#2  0x0000007fecb8dc9c in tflite::Subgraph::ReplaceNodeSubsetsWithDelegateKernels(TfLiteRegistration, TfLiteIntArray const*, TfLiteDelegate*) ()
   from /home/pi/.local/lib/python3.9/site-packages/tflite_runtime/_pywrap_tensorflow_interpreter_wrapper.so
#3  0x0000007fecb8df00 in tflite::Subgraph::ReplaceNodeSubsetsWithDelegateKernels(TfLiteContext*, TfLiteRegistration, TfLiteIntArray const*, TfLiteDelegate*) ()
   from /home/pi/.local/lib/python3.9/site-packages/tflite_runtime/_pywrap_tensorflow_interpreter_wrapper.so
#4  0x0000007fec7cde24 in ?? () from /lib/aarch64-linux-gnu/libedgetpu.so.1
#5  0x0000007fecb8d668 in tflite::Subgraph::ModifyGraphWithDelegate(TfLiteDelegate*) () from /home/pi/.local/lib/python3.9/site-packages/tflite_runtime/_pywrap_tensorflow_interpreter_wrapper.so
#6  0x0000007fecc3d034 in tflite::Interpreter::ModifyGraphWithDelegateImpl(TfLiteDelegate*) () from /home/pi/.local/lib/python3.9/site-packages/tflite_runtime/_pywrap_tensorflow_interpreter_wrapper.so
#7  0x0000007fecb6c8a8 in tflite::interpreter_wrapper::InterpreterWrapper::ModifyGraphWithDelegate(TfLiteDelegate*) () from /home/pi/.local/lib/python3.9/site-packages/tflite_runtime/_pywrap_tensorflow_interpreter_wrapper.so
#8  0x0000007fecb73514 in pybind11::cpp_function::initialize<pybind11_init__pywrap_tensorflow_interpreter_wrapper(pybind11::module_&)::{lambda(tflite::interpreter_wrapper::InterpreterWrapper&, unsigned long)#24}, pybind11::object, tflite::interpreter_wrapper::InterpreterWrapper&, unsigned long, pybind11::name, pybind11::is_method, pybind11::sibling, char [60]>(pybind11_init__pywrap_tensorflow_interpreter_wrapper(pybind11::module_&)::{lambda(tflite::interpreter_wrapper::InterpreterWrapper&, unsigned long)#24}&&, pybind11::object (*)(tflite::interpreter_wrapper::InterpreterWrapper&, unsigned long), pybind11::name const&, pybind11::is_method const&, pybind11::sibling const&, char const (&) [60])::{lambda(pybind11::detail::function_call&)#3}::_FUN(pybind11::detail::function_call&) () from /home/pi/.local/lib/python3.9/site-packages/tflite_runtime/_pywrap_tensorflow_interpreter_wrapper.so
#9  0x0000007fecb822d0 in pybind11::cpp_function::dispatcher(_object*, _object*, _object*) () from /home/pi/.local/lib/python3.9/site-packages/tflite_runtime/_pywrap_tensorflow_interpreter_wrapper.so
#10 0x00000000004cac54 in ?? ()
#11 0x00000000004a5300 in _PyObject_MakeTpCall ()
#12 0x00000000004c6cc8 in ?? ()
#13 0x000000000049c258 in _PyEval_EvalFrameDefault ()
#14 0x00000000004964f8 in ?? ()
#15 0x00000000004b1d20 in _PyFunction_Vectorcall ()
#16 0x00000000004c1d88 in ?? ()
#17 0x00000000004a5180 in _PyObject_MakeTpCall ()
#18 0x000000000049ccf4 in _PyEval_EvalFrameDefault ()
#19 0x00000000004b1a48 in _PyFunction_Vectorcall ()
#20 0x0000000000498064 in _PyEval_EvalFrameDefault ()
#21 0x0000000000496b1c in ?? ()
#22 0x00000000004b1ca0 in _PyFunction_Vectorcall ()
#23 0x0000000000498064 in _PyEval_EvalFrameDefault ()
#24 0x00000000004964f8 in ?? ()
#25 0x0000000000496290 in _PyEval_EvalCodeWithName ()
#26 0x00000000005976fc in PyEval_EvalCode ()
#27 0x00000000005c850c in ?? ()
#28 0x00000000005c2520 in ?? ()
#29 0x00000000005c8458 in ?? ()
#30 0x00000000005c7c38 in PyRun_SimpleFileExFlags ()
#31 0x00000000005b7afc in Py_RunMain ()
#32 0x0000000000587638 in Py_BytesMain ()
#33 0x0000007ff7d00e18 in __libc_start_main (main=0x587538 <_start+56>, argc=10, argv=0x7ffffff118, init=<optimized out>, fini=<optimized out>, rtld_fini=<optimized out>, stack_end=<optimized out>) at ../csu/libc-start.c:308
#34 0x0000000000587534 in _start ()
Backtrace stopped: previous frame identical to this frame (corrupt stack?)

[Skillnoob]

This is now resolved due to the amazing work by feranic in #812 .

[google-coral-bot]

Are you satisfied with the resolution of your issue?
Yes
No