search

microsoft / onnxruntime

ONNX Runtime: cross-platform, high performance ML inferencing and training accelerator
https://onnxruntime.ai
MIT License
13.92k stars 2.81k forks source link

ONNXruntime version 1.18.0 #20877

Open NCVIDevASD opened 2 months ago

NCVIDevASD commented 2 months ago

Describe the issue

the version 1.18.0 is crashing due to an issue

ExecuteKernel] Non-zero status code returned while running GatherElements node. Name:'/dp/flows.7/GatherElements_3' Status Message: /Users/runner/work/1/s/onnxruntime/core/providers/cpu/tensor/gather_elements.cc:154 void onnxruntime::core_impl(const onnxruntime::Tensor , const onnxruntime::Tensor , onnxruntime::Tensor , int64_t, concurrency::ThreadPool ) [Tin = long long] GatherElements op: Out of range value in index tensor

both linux and mac have this issue , downgrading to 1.17 fix ,

To reproduce

just used piper , it is using the latest version of onnxruntime ,

Urgency

not so urgent but cause some troubleshooting on the server side to downgrade all onnxruntime

Platform

Linux

OS Version

22.04

ONNX Runtime Installation

Released Package

ONNX Runtime Version or Commit ID

1.18

ONNX Runtime API

Python

Architecture

X64

Execution Provider

Default CPU

Execution Provider Library Version

No response

Betty1202 commented 2 months ago

I have the same issue, and I can't fix it by downgrading onnxruntime. I even try 1.16, 1.15, 1.14, and all of them have this issue. BTW, I can successfully run the script a few days ago.

jywu-msft commented 2 months ago

can you provide more details/repro instructions?

jywu-msft commented 2 months ago

@Betty1202 which script are you referring to?

Betty1202 commented 2 months ago

@jywu-msft I have solved it, thanks!

cnndabbler commented 2 months ago

@Betty1202 how did you solve it? thx

Betty1202 commented 2 months ago

@cnndabbler I changed some of the tensors from int64 to int when calculating the metric on the prediction. I don't know why this would have an effect on onnxruntime though, but it's amazingly fixed!

github-actions[bot] commented 1 month ago

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.