Does random seed have any effect on the order of data loader when shuffle is set to False?

I define the seeds in the following way:

torch.manual_seed(rs)
torch.cuda.manual_seed(rs)
torch.cuda.manual_seed_all(rs)
torch.backends.cudnn.deterministic=True
torch.backends.cudnn.benchmark = False
np.random.seed(rs)
random.seed(rs)

And the data loader as:

testloader = DataLoader(testset, batch_size=16, shuffle=False, num_workers=2, pin_memory=True)

Will the order of the samples be different if I change the random seed rs ?

Would setting number of workers to 0 be helpful in the case each background process can lead to different order of samples ?

The order should be deterministic and not affected by the seeding if no suffling is used. You could verify it by iterating your DataLoader and comparing e.g. the .abs().sum() values of each sample.