PyTorch failure after onnx update

To: Debian Deep Learning Team <debian-ai@lists.debian.org>
Subject: PyTorch failure after onnx update
From: Gard Spreemann <gspr@nonempty.org>
Date: Mon, 27 Jun 2022 13:31:09 +0200
Message-id: <[🔎] 87zghye3wz.fsf@nonempty.org>

Hi,

After a recent upgrade of libonnx1 to 1.12.0-1, PyTorch started failing:

 ImportError: /lib/x86_64-linux-gnu/libtorch_cpu.so.1.8: undefined symbol: _ZN4onnx12optimization8OptimizeERKNS_10ModelProtoERKSt6vectorINSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEESaISA_EE

Are we just looking at a need to rebuild, or something deeper?


 Best,
 Gard

Attachment: signature.asc
Description: PGP signature

Reply to:

Follow-Ups:
- Re: PyTorch failure after onnx update
  - From: "M. Zhou" <lumin@debian.org>

Prev by Date: tensorpipe_0.0~git20220513.bb1473a-2_source.changes ACCEPTED into unstable
Next by Date: Re: PyTorch failure after onnx update
Previous by thread: tensorpipe_0.0~git20220513.bb1473a-2_source.changes ACCEPTED into unstable
Next by thread: Re: PyTorch failure after onnx update
Index(es):
- Date
- Thread