How to pass inputs for my triton model using tritionclient python package

Gaming and Visualization Technologies General Topics and Other SDKs
1

My triton model config.pbtxt file looks like below. How can I pass inputs and outputs using tritonclient and perform an infer request.

name: “cifar10”
platform: “tensorflow_savedmodel”
max_batch_size: 10000
input [
{
name: “input_1”
data_type: TYPE_FP32
dims: [ 32, 32, 3 ]
}
]
output [
{
name: “fc10”
data_type: TYPE_FP32
dims: [ 10 ]
}
]

Any help is appreciated as I am very new to using tritonclient python package please

2

Hi there @magr90900 and welcome to the NVIDIA developer forums!

The general Triton support has moved to Github issues a while ago, which means questions specific to that Github will get better exposure and replies over there.

If you want to get more general advice on inference topics, you might want to check out our DL specific categories Deep Learning (Training & Inference) - NVIDIA Developer Forums

I hope that helps!