[IE-MYRIAD] Chromium crashed when running Facial Landmark

[Christywl]

Test Env: Chromium Version: nightly build 79.0.3917.0 (74a844d) Platform: Linux

Expected Result: The tests should pass.

Actual Result: Plug Intel Neural Compute Stick 2, it throws an error in the first time: [27833:27954:0612/161029.865310:ERROR:execution_impl_ie.cc(97)] [IE] exception Can not init Myriad device: NC_ERROR

unplug and plug Intel Neural Compute Stick 2 again, Chromium crashed:

` [27833:27954:0612/161345.397646:ERROR:execution_impl_ie.cc(192)] [IE] exception Failed to queue inference: NC_ERROR Received signal 11 SEGV_MAPERR 000000000008

0 0x55d45709ecf9

1 0x55d457002373

2 0x55d45709e881

3 0x7f4ba64d9390

4 0x55d455a88638

5 0x55d455a750a0

6 0x55d455a6eb36

7 0x55d4571c4b8b

8 0x55d4571ca61e

9 0x55d4571c9e98

10 0x55d4571c29a6

11 0x55d4571c337e

12 0x55d4571d88de

13 0x55d45704d5bf

14 0x55d457064f32

15 0x55d457064a69

16 0x55d4570ae80b

17 0x55d4570646be

18 0x55d45706e883

19 0x55d45706e6f4

20 0x55d4570aefde

21 0x7f4ba64cf6ba start_thread

22 0x7f4ba044641d clone

r8: 0000000000000000 r9: 0000000000000000 r10: 0000000000004000 r11: 0000000000000000 r12: 0000000000000001 r13: 000005769294c000 r14: 00007f4b9093eca0 r15: 0000057690e82ca0 di: 0000000000000000 si: 0000000000000000 bp: 00007f4b9093edf0 bx: 00000576913f8ed0 dx: 0000057690e82c40 ax: 00007f4b9093ee50 cx: fffffffd515d5b54 sp: 00007f4b9093ec40 ip: 000055d455a88638 efl: 0000000000010246 cgf: 002b000000000033 erf: 0000000000000004 trp: 000000000000000e msk: 0000000000000000 cr2: 0000000000000008 [end of stack trace] Calling _exit(1). Core file will not be generated. [28176:28176:0612/161345.522200:ERROR:buffer_manager.cc(488)] [.DisplayCompositor]GL ERROR :GL_INVALID_OPERATION : glBufferData: <- error from previous GL command `

How to Reproduce:

1. Launch chromium with /usr/bin/chromium-browser-unstable --no-sandbox --use-inference-engine and visit https://intel.github.io/webml-polyfill/examples/facial_landmark_detection
2. Plug Intel Neural Compute Stick 2
3. Select LOW_POWER
4. Select SimpleCNN(TFLite)

5. Select one model:

   • SSD MobileNet v1(TFLite)
   • SSD MobileNet v1(TFLite)
   • SSDLite MobileNet v2(TFLite)