fix autoencoder memory tests #38572

Job	Run time
check_code_quality	33s
check_repository_consistency	27s
Hub tests for models, schedulers, and pipelines	2m 29s
LoRA tests with PEFT main	6m 25s
Fast PyTorch Pipeline CPU tests	15m 57s
PyTorch Example CPU tests	19m 32s
Fast PyTorch Models & Schedulers CPU tests	7m 14s
	52m 37s

Provide feedback