Inference subgraph to sample accumulator pipeline

Hi there,

I want to set up a sample accumulator to monitor what objects have been detected and trigger python GPIO code once the app sees a particular object. I’m having trouble setting up the sample accumulator node in my app. This is what I have added to an otherwise functional app:

In config:

"sample_accumulator": {

“SampleAccumulator”: {
“sample_buffer_size”: 256,
}
},

In nodes:

  {
    "name": "sample_accumulator",
    "components": [
      {
        "name": "message_ledger",
        "type": "isaac::alice::MessageLedger"
      },
      {
        "name": "SampleAccumulator",
        "type": "isaac::ml::SampleAccumulator"
      }
    ]
  },

In edges:

  {
    "source": "inference.subgraph/interface/detections",
    "target": "sample_accumulator/SampleAccumulator/samples"
  }

Please let me know if I’m missing something. Also, the documentation says

“Sample Accumulator: Takes in the training pairs (image tensor and detection label tensor) as a TensorListProto and stores that in a buffer. This codelet is bound to the python script such that the training script can directly sample from this buffer with the acquire_samples() function. The acquire_samples() function converts the TensorListProto into a list of numpy arrays with corresponding dimensions and pass that to Python.”— https://docs.nvidia.com/isaac/isaac/packages/yolo/doc/yolo.html#codelets

I’m wondering where is the python script that this codelet is bound to? Do I need to set this up? There is no output listed for this node so I’m a bit confused. Thanks in advance for any help you can offer.

Hi,

Are you facing some issue when using our isaac SDK?
If yes, I will move your topic to our isaac board to get more support.

Thanks.

Hi AastaLLL,

Yes, thank you. That would be great.