Tanmay Verma

We are also seeing a hang-up in our fuzz test when moving from boofuzz 0.3.0 to 0.4.0. See [here](https://github.com/triton-inference-server/server/blob/main/qa/L0_http_fuzz/fuzztest.py).

@alexhcheng Can you move your PR from official repository to contrib? We would not want to officially support C#. https://github.com/triton-inference-server/contrib

The javascript contribution came from someone internal to our company and there was an ask to host it. Any clients we add needs to be tested in our CI. We...

@rarzumanyan Can you describe how your changes work? For device=gpu, what is the input_values for set_shared_memory_region? I would suggest adding an example demonstrating the DtoD copies similar to [simple_http_cudasm_client](https://github.com/triton-inference-server/client/blob/main/src/python/examples/simple_http_cudashm_client.py). You...

> {'name': 'test_1_values', 'type': {'tensorType': {'elemType': 1, 'shape': {'dim': [{}]}}}} Triton does not support scalar tensors. You would have modify the network's tensors to shape [1].

Yes.. as the shape of the tensors are empty it looks like they are scalars. Modifying the vectors to non-scalar, either fixed or variable shape can help you successfully loading...

Interesting... We do have tests for dynamic-shapes I/O. Can you share a simple model that can be used for reproducing the issue? We will investigate further.

Unfortunately, I am getting the following error when trying to load the shared model. ``` E1020 00:23:50.969259 112 model_repository_manager.cc:1186] **failed to load 'test_model' version 1: Internal: onnx runtime error 1:...

@askhade Do you have any insights into the error? The error does make it look like an issue with ONNXRT/openvino integration issue but the model seems to work with python...

@mayani-nv BERT is data-sensitive model. perf_analyzer by default use random data to fill in tensors and model might not like that. You should be able to probably get it working...