Part Number: PROCESSOR-SDK-AM57X
Other Parts Discussed in Thread: AM5728, AM5749
Hi,
I have trained a mobilenet based SSD on VOC dataset, the framework I used is caffe-jacinto(without sparse trainning step). The caffemodel have been successfully imported to what TIDL surpports,and the network have been seperated to two layergroups just as the ssd models in ti-processor-sdk-linux-am57xx-evm-05.01.00.11/linux-devkit/sysroots/armv7ahf-neon-linux-gnueabi/usr/share/ti/tidl/examples/test/testvecs/config/tidl_models.
I tried to use the application in ti-processor-sdk-linux-am57xx-evm-05.01.00.11/linux-devkit/sysroots/armv7ahf-neon-linux-gnueabi/usr/share/ti/tidl/examples/ssd_multibox, the default configuration wroks just fine on AM57xEVM and I can get the correct output. But when I change the confguration file to point to my own model bin file, the app just can't execut successfully. I have add some print impromation,and found that eop->ProcessFrameWait() can't get a process finished signal from eop->ProcessFrameStartAsync(), so the app just stuck in process.
I don't know why this happened.mabye there are wrongs in my imported model or other reasons have caused this problem?
best regards,
