Yue Wu

### Describe the bug In the diffusers [basic training tutorial](https://huggingface.co/docs/diffusers/tutorials/basic_training), setting the manual seed argument (`generator=torch.manual_seed(config.seed)`) in the pipeline call inside evaluate() function rewinds the dataloader shuffling, leading to overfitting...

# What does this PR do? In the diffusers basic training tutorial, setting the manual seed argument (generator=torch.manual_seed(config.seed)) in the pipeline call inside the evaluate() function rewinds the random state...