Skip to content
Submit incident
AII-1001-87467
ReportedReported by fewer than three independent publishers, with no official finding.

PyTorch torch.jit.jit_module_from_flatbuffer segmentation fault

A segmentation fault was reported in PyTorch's torch.jit.jit_module_from_flatbuffer function when loading and executing a JIT-scripted model in version 2.6.0. PyTorch closed the issue as not planned, citing deprecation of TorchScript in favor of torch.export.

22 March 2025 · PyTorch · Reliability and hallucination
22 March 2025

Drafted by Raidu Incident Detection, an automated system.

AII-1001-87467View source ↗
Quick facts
Organization
PyTorch (subject)
Category
Reliability and hallucination
Source
github.com
Date
22 March 2025
Drafted by
Raidu Incident Detection

What happened

A segmentation fault occurred when loading a JIT module from a flatbuffer file using torch.jit.jit_module_from_flatbuffer in PyTorch version 2.6.0.

The issue was triggered when a scripted model containing a Sequential module with Linear, BatchNorm2d, and ReLU layers was saved to flatbuffer format and then loaded and executed with sample input.

Index analysis

The segmentation fault represents a memory corruption vulnerability that could lead to denial of service or potential code execution.

PyTorch closed the issue as not planned on August 7, 2026, noting that TorchScript has been deprecated in favor of torch.export as of version 2.11.

Classification

Organization
PyTorch (subject)
Category
Reliability and hallucination
Harm status
Actual harm
Intent
Unintentional
Timing
Unknown
Consequences
None recorded
AI systems
PyTorch

Sources and evidence

This record was researched and written by the Index, based on the following original reporting.

github.com
https://github.com/pytorch/pytorch/issues/149800