[Performance] Memory leak in Python

[franz101]

Describe the issue

After deleting a session in python, the memory is still allocated (in the onnx runtime). Over time if multiple sessions are created this leaks to the memory getting overloaded. I attached a simple POC in colab

To reproduce

# RAM % before: 6.8
sessions = [ort.InferenceSession(model_path) for _ in range(500)]
del sessions
print(f"RAM % after: {psutil.virtual_memory().percent}")
# RAM % after: 63.6 (expected below 8%)

Further notebook:
https://colab.research.google.com/drive/1Q8OrYxt2PKGvXiWOw-LRjc7XV5mN7u1u#scrollTo=9UWHzgu76FaF

Urgency

This currently is crashing our production backend written in Python.

Platform

Linux

OS Version

Ubuntu 18.04

ONNX Runtime Installation

Released Package

ONNX Runtime Version or Commit ID

1.16.3

ONNX Runtime API

Python

Architecture

X64

Execution Provider

Default CPU

Execution Provider Library Version

No response

Model File

https://colab.research.google.com/drive/1Q8OrYxt2PKGvXiWOw-LRjc7XV5mN7u1u#scrollTo=9UWHzgu76FaF

Is this a quantized model?

Unknown

### Tasks

[franz101]

Tagging:
@snnn, @TMVector and @jambayk

[franz101]

Interestingly it worked in version onnx 1.8.1 and onnx runtime 1.6.0 (onnx/onnx#3293)
could be related to (#16117)

[yuslepukhin]

https://stackoverflow.com/questions/21053380/what-does-del-do-exactly

[ShuaiShao93]

https://stackoverflow.com/questions/21053380/what-does-del-do-exactly

If you look at the notebook, there is gc.collect after del

[github-actions]

This issue has been automatically marked as stale due to inactivity and will be closed in 30 days if no further activity occurs. If further support is needed, please provide an update and/or more details.

[github-actions]

This issue has been automatically closed due to inactivity. Please reactivate if further support is needed.