search

microsoft / DeepSpeed

DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.
https://www.deepspeed.ai/
Apache License 2.0
33.63k stars 3.95k forks source link

Add compile backend arg for test_set_compiler_fn #5611

Closed vshekhawat-hlab closed 3 weeks ago

vshekhawat-hlab commented 3 weeks ago

Fix the test_set_compiler_fn UT, current test_set_compiler_fn uses inductor as a backend, which is incompatible with HPU. Added backend arg to torch.compile.

tohtana commented 3 weeks ago

Hi, @vshekhawat-hlab, we are considering to remove set_compiler_fn and its test in #5581. I think we can merge it within this week. Can you wait for a while?

vshekhawat-hlab commented 3 weeks ago

Thanks @tohtana for information.

I am closing this.