OOM errors · jax-ml jax · Discussion #10543 · GitHub

zohimchandani
May 3, 2022

What does one substitute for platform in XLA_PYTHON_CLIENT_ALLOCATOR=platform?

Replies: 1 comment

jakevdp
May 3, 2022
Maintainer

The valid options are ["default", "platform", "bfc", "cuda_async"]. See https://github.com/tensorflow/tensorflow/blob/77004536aff979d029dccd12880dfa9179933442/tensorflow/compiler/xla/python/xla_client.py#L73

I don't know of any detailed documentation for this setting.

0 replies

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment