how to release tpu memory after del diffusers pipeline

[ghost]

❓ Questions and Help

i create a pipeline
`
pipeline = DiffusionPipeline.from_pretrained("stable-diffusion-v1-5/stable-diffusion-v1-5", torch_dtype=torch.bfloat16).to(torch_xla.core.xla_model.xla_device())

pipeline.to('cpu')

pipeline = StableAudioPipeline.from_pretrained("stabilityai/stable-audio-open-1.0", torch_dtype=torch.bfloat16).to(torch_xla.core.xla_model.xla_device()) #which cause tpu memory problem
`
i want to ask how to release tpu memory. is there any tpu version of torch.cuda.empty_cache()？

[radna0]

See #8520 (comment)