search

intel / intel-extension-for-tensorflow

Intel® Extension for TensorFlow*
Other
323 stars 40 forks source link

AI-Benchmark unable to complete on WSL2 #53

Closed FabianRamirez03 closed 8 months ago

FabianRamirez03 commented 1 year ago

Hi, I'm trying to run AI benchamark on a ARC A750 but is unable to complete on step 11/19.

This is the code of the benchmark that I'm trying to run:

from ai_benchmark import AIBenchmark

benchmark = AIBenchmark(use_CPU=None)

results = benchmark.run()

This is the error message that I got in the 11/19 test:

11/19. ResNet-DPED

2023-11-18 23:30:58.863034: I tensorflow/core/common_runtime/pluggable_device/pluggable_device_factory.cc:306] Could not identify NUMA node of platform XPU ID 0, defaulting to 0. Your kernel may not have been built with NUMA support.
2023-11-18 23:30:58.863314: I tensorflow/core/common_runtime/pluggable_device/pluggable_device_factory.cc:272] Created TensorFlow device (/job:localhost/replica:0/task:0/device:XPU:0 with 0 MB memory) -> physical PluggableDevice (device: 0, name: XPU, pci bus id: <undefined>)
2023-11-18 23:30:59.656162: F ./itex/core/utils/tensor_format.h:432] Check failed: index >= 0 && index < num_total_dims Invalid index from the dimension: 3, 0, C
If you need help, create an issue at https://github.com/intel/intel-extension-for-tensorflow/issues
Aborted

I attach the full logs. ai-benchmark-logs.txt

I followed this wiki in order to setup the WSL2 environment.

I know the intel_extension_for_tensorflow is working because it reflects in the GPU usage. Each wave of usage is an attempt of usage of the scrpi, in the image are two attempts.. When I tried to use it before without the extension, this doesn't happened. image

Let me know if you need anymore details, Regards,

-Fabian

jianyizh commented 1 year ago

Hi, we have fixed this error https://github.com/intel/intel-extension-for-tensorflow/commit/6a01f118fd847a62d49d3536a18d491ef3ba8238. Please have a try.