nodejs + tensorrt :: how to set execution provider options?

[ogolovanov]

Describe the issue

Hi.

I unable to find any examples on the web - how to set provider options for TensorRT via nodejs.
At the same time there are examples for C++/python/JAVA.
https://onnxruntime.ai/docs/execution-providers/TensorRT-ExecutionProvider.html#samples

Setting provider options via ENV also does not work.

export ORT_TENSORRT_ENGINE_CACHE_ENABLE=1
export ORT_TENSORRT_CACHE_PATH=/tmp/trt_cache/
node test.js

test.js creates session like that:
const model = await ort.InferenceSession.create(modelBuffer, {"executionProviders": ["tensorrt"], "logSeverityLevel": 0})

Everyting works ok, but there is nothing at /tmp/trt_cache/.
I see that nodejs binding simply ignores any ENV variables, related to TensorRT.

Same situation with "onnxruntime_perf_test" ( = nothing at /tmp/trt_cache/ ):

export ORT_TENSORRT_ENGINE_CACHE_ENABLE=1
export ORT_TENSORRT_CACHE_PATH=/tmp/trt_cache/
./onnxruntime_perf_test -r 1 -e tensorrt -i "trt_fp16_enable|true" /root/www/model.trt.onnx

I don't understand something, or it is really not possible to set TensorRT provider options via nodejs?
Why ORT_TENSORRT_* ENV variables ignored?

Thanks.

To reproduce

---

Urgency

not urgent

Platform

Linux

OS Version

Ubuntu 22.04

ONNX Runtime Installation

Built from Source

ONNX Runtime Version or Commit ID

1.17.0

ONNX Runtime API

JavaScript

Architecture

X64

Execution Provider

TensorRT

Execution Provider Library Version

TensorRT 8.6.1.6-1+cuda12.0

[github-actions]

This issue has been automatically marked as stale due to inactivity and will be closed in 7 days if no further activity occurs. If further support is needed, please provide an update and/or more details.

[fr-an-k]

Is tensorrt supported at all on the nodejs backend? It would be good to add a tensorrt row to the table of supported backends in the documentation, even if it is not supported.