AML_NN_SDK的demo报Error:not support this driver

Which Khadas SBC do you use?

VIM3

Which system do you use? Android, Ubuntu, OOWOW or others?

Ubuntu

Which version of system do you use? Khadas official images, self built images, or others?

Khadas official images

Please describe your issue below:

在使用SDK API的过程在看到在 Amlogic-NN/AML_NN_SDK (github.com) 中有给出很多demo,便尝试着去运行demo。但是按照https://github.com/Amlogic-NN/AML_NN_SDK/blob/master/Linux/examples/readme.txt 的步骤进行拷贝到板子上运行不了。

Post a console log of your issue below:

demo编译结果:
-- Found Threads: TRUE  
-- Configuring done
-- Generating done
-- Build files have been written to: /mnt/hgfs/security_camera/khadas/AML_NN_SDK-master/AML_NN_SDK-master/Linux/examples/build
Scanning dependencies of target aml_image_cls_64
[ 33%] Building CXX object demo/image_classify/CMakeFiles/aml_image_cls_64.dir/main.cpp.o
[ 66%] Building C object demo/image_classify/CMakeFiles/aml_image_cls_64.dir/postprocess_classify.c.o
[100%] Linking CXX executable ../build/aml_image_cls_64
[100%] Built target aml_image_cls_64

板子上报错:
root@Khadas:/home/khadas/sdk_demo# ls
224x224x3.jpeg  aml_image_cls_64  image_classify_88_A311D.nb
root@Khadas:/home/khadas/sdk_demo# ./aml_image_cls_64 image_classify_88_A311D.nb 224x224x3.jpeg 
-bash: ./aml_image_cls_64: Permission denied
root@Khadas:/home/khadas/sdk_demo# chmod 777 aml_image_cls_64 
root@Khadas:/home/khadas/sdk_demo# ./aml_image_cls_64 image_classify_88_A311D.nb 224x224x3.jpeg 
50
the input type should be 224*224*3
E NN_SDK:[aml_module_create:82]Error:not support this driver. 
E NN_SDK:[aml_module_create:91]Error:aml_unify_init fail, context is null. 
amlnn_init is fail
E NN_SDK:[aml_module_input_set:108]Error:aml_module_input_set fail, context is null. 
aml_module_input_set error
E NN_SDK:[aml_module_output_get:161]Error:aml_module_output_get fail, context is null. 
aml_module_output_get error
E NN_SDK:[aml_module_destroy:289]Error:aml_module_destroy fail, context is null 
aml_module_destroy error

gdb进行debug发现在这里报错,qcontext是NULL的
	config.modelType = TENSORFLOW;
	qcontext = aml_module_create(&config);
	if (qcontext == NULL)
	{
		printf("amlnn_init is fail\n");
		return NULL;
我能想到的原因就是这些demo的SDK版本与我板子上不同。板子上是DDK6.4.8.7的
但是这个问题要怎么解决呢,@Frank 之前说现在可以直接在板子上编译了,但是这个SDK API要怎么导入到我的代码中?

搞着搞着报另一个错误 :sweat_smile: :sweat_smile:
101
the input type should be 2242243
E NN_SDK: [amlnn_init:850]Error:network verifygraph fail
E NN_SDK: [amlnn_init:896]Error:amlnn_init fail
E NN_SDK: [aml_module_create:70]Error:amlnn_init is fail, context is null
amlnn_init is fail
E NN_SDK: [aml_module_input_set:84]Error:aml_module_input_set is fail, context is null
aml_module_input_set error
E NN_SDK: [aml_module_output_get:119]Error:aml_module_output_get is fail, context is null
aml_module_output_get error
E NN_SDK: [aml_module_destroy:266]Error:aml_module_destroy is fail, context is null
aml_module_destroy error