Lora does not support module_to_save()

### Describe the bug

When we set
```
LoraConfig(
r=args.rank,
lora_alpha=args.lora_alpha,
init_lora_weights=True,
target_modules=["to_k", "to_q", "to_v", "to_out.0"],
modules_to_save=["other"],
)
```
if `modules_to_save` is not none, then we load the adapter,`diffusers/src/diffusers/loaders/lora_pipeline.py` line 2681 `is_correct_format = all("lora" in key for key in state_dict.keys())` will be `False`

### Reproduction

See Descrination

### Logs

_No response_

### System Info

diffusers 0.32.0

### Who can help?

_No response_

Start in diffusers/src/diffusers/loaders/lora_pipeline.py at line 2681 and inspect how the state_dict format check behaves when LoraConfig includes modules_to_save. Reproduce the reported configuration with diffusers 0.32.0, then verify that loading an adapter with modules_to_save is accepted without the incorrect format result.

Scritto dal modello di indicizzazione a partire dal testo della issue.