Getting errors with STM32Cube AI Studio

[emmanuel_]

Dear all,

With the new STM32Cube AI Studio tool, I am getting the following errors (see attached log file). I am using it with the kit STM32N6570-DK.

I tried the same model as in the tutorial (https://www.youtube.com/watch?v=3zyIpB5o9zQ). I am getting the same type of error with other models as well.

Thank you for your help. I am looking forward to being able to use this new tool.

-------------------------------------------------------------------------------------------------------------------------------------------------

C:/Users/papam/.stm32cubeaistudio/workspace/demo_face_detection/.ai/run/run-1/Middlewares/ST/AI/Validation/Src/ai_wrapper_ATON.c:386:10: warning: suggest parentheses around assignment used as truth value [-Wparentheses]

386 | while (aton_buf = npu_get_input_buffers_info(instance, nn_info->n_inputs))
| ^~~~~~~~

C:/Users/papam/.stm32cubeaistudio/workspace/demo_face_detection/.ai/run/run-1/Middlewares/ST/AI/Validation/Src/ai_wrapper_ATON.c:393:10: warning: suggest parentheses around assignment used as truth value [-Wparentheses]

393 | while (aton_buf = npu_get_output_buffers_info(instance, nn_info->n_outputs))
| ^~~~~~~~

In file included from C:/Users/papam/.stm32cubeaistudio/workspace/demo_face_detection/.ai/run/run-1/Middlewares/ST/AI/Npu/ll_aton_profiler.c:27:

C:/Users/papam/.stm32cubeaistudio/workspace/demo_face_detection/.ai/run/run-1/Middlewares/ST/AI/Npu/ll_aton_profiler.c: In function 'LL_ATON_LIB_ConvInteger':

C:/Users/papam/.stm32cubeaistudio/workspace/demo_face_detection/.ai/run/run-1/Middlewares/ST/AI/Npu/ll_aton_profiler.c:206:31: warning: format '%d' expects argument of type 'int', but argument 4 has type 'int32_t' {aka 'long int'} [-Wformat=]

206 | LL_ATON_PROFILER_PRINTF("%s k=%d %d scale=%g\n", conv_name, k, maxmax_k, kern->scale[k]);
| ^~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~
| |
| int32_t {aka long int}

C:/Users/papam/.stm32cubeaistudio/workspace/demo_face_detection/.ai/run/run-1/Middlewares/ST/AI/Npu/ll_aton_util.h:43:46: note: in definition of macro 'LL_ATON_PROFILER_PRINTF'
43 | #define LL_ATON_PROFILER_PRINTF(...) printf(__VA_ARGS__)
| ^~~~~~~~~~~

C:/Users/papam/.stm32cubeaistudio/workspace/demo_face_detection/.ai/run/run-1/Middlewares/ST/AI/Npu/ll_aton_profiler.c:206:41: note: format string is defined here
206 | LL_ATON_PROFILER_PRINTF("%s k=%d %d scale=%g\n", conv_name, k, maxmax_k, kern->scale[k]);
| ~^
| |
| int

| %ld
C:/Users/papam/.stm32cubeaistudio/workspace/demo_face_detection/.ai/run/run-1/Middlewares/ST/AI/Npu/ll_aton_profiler.c:282:33: warning: format '%d' expects argument of type 'int', but argument 4 has type 'int32_t' {aka 'long int'} [-Wformat=]
282 | LL_ATON_PROFILER_FPRINTF(f, "%s %d\n", conv_name, maxmax); // log((float)maxmax) / log(2.0));
| ^~~~~~~~~ ~~~~~~
| |

| int32_t {aka long int}

[emmanuel_]

After investigating ,  I installed the new version  of  SEdgeAI3\3.0\  and uptdate the Local path in Settings

cheers

Emmanuel

Local Path  ST Edge AI Core

...ST\STEdgeAI3\3.0\Utilities\windows\stedgeai.exe

[hamitiya]

Hello @emmanuel_ 

Are you reproducing this issue for every models you tried so far? Did it also fail with ST Edge AI Core 3.0?

Based on the logs, it seems that the execution worked fine but it failed to get results.

I will try to replicate the issue and keep you in touch.

Best regards,

Yanis

[hamitiya · Accepted solution]

I reproduced the issue with 2.2. 

A fix will be ready for the next update.

For now, please use the 3.0 since I was not able to reproduce the issue with this version.

Best regards,

Yanis

[emmanuel_]

Hi Yanis ,

Thank's a lot .

cheers

Emmanuel