error when switch to workflow TorchCompileVAE

[15053285989]

hi, I tried run workflow A(with TorchCompileVAE) first,
then switch to workflow B(without TorchCompileVAE),
and then switch back to workflow A(with TorchCompileVAE),
I got error as follows:

07:38:46 - !!! Exception during processing !!! Failed to compile model
07:38:46 - Traceback (most recent call last):
  File "/mnt/app/custom_nodes/ComfyUI-KJNodes/nodes/model_optimization_nodes.py", line 346, in compile
    vae.first_stage_model.encoder = torch.compile(vae.first_stage_model.encoder, mode=mode, fullgraph=fullgraph, backend=backend)
  File "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py", line 1984, in __setattr__
    raise TypeError(
TypeError: cannot assign 'comfy.ldm.modules.diffusionmodules.model.Encoder.forward' as child module 'encoder' (torch.nn.Module or None expected)

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/mnt/app/execution.py", line 328, in execute
    output_data, output_ui, has_subgraph = get_output_data(obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb)
  File "/mnt/app/execution.py", line 203, in get_output_data
    return_values = _map_node_over_list(obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb)
  File "/mnt/app/execution.py", line 174, in _map_node_over_list
    process_inputs(input_dict, i)
  File "/mnt/app/execution.py", line 163, in process_inputs
    results.append(getattr(obj, func)(**inputs))
  File "/mnt/app/custom_nodes/ComfyUI-KJNodes/nodes/model_optimization_nodes.py", line 349, in compile
    raise RuntimeError("Failed to compile model")
RuntimeError: Failed to compile model

[15053285989]

the reason is that compile flag is in node instanc, and when node is recycled, vae will be compiled twice.

            if not self._compiled_encoder:
                try:
                    vae.first_stage_model.encoder = torch.compile(vae.first_stage_model.encoder, mode=mode, fullgraph=fullgraph, backend=backend)
                    self._compiled_encoder = True
                except Exception as e:
                    logging.exception(e)