System crash or shutdown in Resnet50 inference

[merazlab]

Thanks Intel AI model team,

I run Image recognition Resnet-50 inference model on Intel ICX machine for batch size=1, numa_core_per_instance=1, precision=FP32 , then after system shutdown automatically, Why?, but some other combination works well.

https://github.com/IntelAI/models/blob/master/benchmarks/launch_benchmark.py

[MateuszGrabuszynski]

Hello, I recently encountered issues with RN50 too. The solution for me was to ensure I am aligned with intel-extension-for-tensorflow Compatibility table from here https://github.com/intel/intel-extension-for-tensorflow#compatibility-table. For more info, check issue linked above.
(EDIT: I have not checked the opening date here...)

[sramakintel]

@merazlab can you try our latest optimizations for resnet50 to see if the issue is resolved? https://www.intel.com/content/www/us/en/developer/articles/containers/cpu-reference-model-containers.html