• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 97 subscribers
  • Views 484 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TDA4VM: TDA4: Jacinto 7 C71 throw a problem, when use the jacinto-ai-devkit net work

user6480120
Prodigy 80 points

Part Number: TDA4VM

Hi:

       I got a problem use this network, when I just adjust the weight and number of classification.

My board is TDA4, the C71 throw an error

“[C7x_1 ] 64.731472 s: VX_ZONE_ERROR:[tivxAlgiVisionCreate:331] Calling ialg.algAlloc failed with status = -1115”

It lead verify graph fail

64.731630 s: VX_ZONE_ERROR:[ownContextSendCmd:770] Command ack message returned failure cmd_status: -1
64.731649 s: VX_ZONE_ERROR:[ownContextSendCmd:806] tivxEventWait() failed.
64.731671 s: VX_ZONE_ERROR:[ownNodeKernelInit:526] Target kernel, TIVX_CMD_NODE_CREATE failed
64.731803 s: VX_ZONE_ERROR:[ownGraphNodeKernelInit:583] kernel init for node 2, kernel com.ti.tidl ... faile!
64.731814 s: VX_ZONE_ERROR:[vxVerifyGraph:2010] Node kernel init failed
64.731820 s: VX_ZONE_ERROR:[vxVerifyGraph:2064] Graph verify failed

   The demo I use is based on tidl_app_od. I don’t know why it report these problems.

 

Best regards

  • 0
    Prodigy 80 points

    Hi

         

    Firstly, I retrained the ourself model  based on the  JDetNet caffe model(512*512),just modify the class number , as the deploy.

    The deploy, solver and weights download from the github (ttps://github.com/TexasInstruments/jacinto-ai-devkit)

    Then I import the model by the JDetNet_ft.txt on PC simulation ,the detection result is fine ,almost same with the desktop(without the ti sdk enviroment)

     However, I test the model.bin file on EVM , and the C71 throw an error.

    “[C7x_1 ] 64.731472 s: VX_ZONE_ERROR:[tivxAlgiVisionCreate:331] Calling ialg.algAlloc failed with status = -1115”

    It lead verify graph fail

    64.731630 s: VX_ZONE_ERROR:[ownContextSendCmd:770] Command ack message returned failure cmd_status: -1
    64.731649 s: VX_ZONE_ERROR:[ownContextSendCmd:806] tivxEventWait() failed.
    64.731671 s: VX_ZONE_ERROR:[ownNodeKernelInit:526] Target kernel, TIVX_CMD_NODE_CREATE failed
    64.731803 s: VX_ZONE_ERROR:[ownGraphNodeKernelInit:583] kernel init for node 2, kernel com.ti.tidl ... faile!
    64.731814 s: VX_ZONE_ERROR:[vxVerifyGraph:2010] Node kernel init failed
    64.731820 s: VX_ZONE_ERROR:[vxVerifyGraph:2064] Graph verify failed.

    The  attachment is the model I used.NetModel.zip

  • 0
    Prodigy 80 points

    Hi

         

    Firstly, I retrained the ourself model  based on the  JDetNet caffe model(512*512),just modify the class number , as the deploy.

    The deploy, solver and weights download from the github (ttps://github.com/TexasInstruments/jacinto-ai-devkit)

    Then I import the model by the JDetNet_ft.txt on PC simulation ,the detection result is fine ,almost same with the desktop(without the ti sdk enviroment)

     However, I test the model.bin file on EVM , and the C71 throw an error.

    “[C7x_1 ] 64.731472 s: VX_ZONE_ERROR:[tivxAlgiVisionCreate:331] Calling ialg.algAlloc failed with status = -1115”

    It lead verify graph fail

    64.731630 s: VX_ZONE_ERROR:[ownContextSendCmd:770] Command ack message returned failure cmd_status: -1
    64.731649 s: VX_ZONE_ERROR:[ownContextSendCmd:806] tivxEventWait() failed.
    64.731671 s: VX_ZONE_ERROR:[ownNodeKernelInit:526] Target kernel, TIVX_CMD_NODE_CREATE failed
    64.731803 s: VX_ZONE_ERROR:[ownGraphNodeKernelInit:583] kernel init for node 2, kernel com.ti.tidl ... faile!
    64.731814 s: VX_ZONE_ERROR:[vxVerifyGraph:2010] Node kernel init failed
    64.731820 s: VX_ZONE_ERROR:[vxVerifyGraph:2064] Graph verify failed.

    The  attachment is the model I used.2772.NetModel.zip

  • 0 in reply to user6480120
    TI__Mastermind 22145 points

    Hi,

    Can you share the complete import log of this model.

    The error message (-1115) TIDL_E_QUANT_STATS_NOT_AVAILABALE means that the import step did not complete the quantization claibrartion propoerly

  • 0 in reply to kumar.desappan
    Prodigy 80 points

    Hi :

    You're right. I found the mistake I made. The model had not finished learning, when I stopped it . 

    Thank you for your reply. I will set the issue status as resolved.