SPOS hardware-aware nas multi_trial.py don't work

[GeorgeW9527]

Describe the issue:

Environment: ubutu

• NNI version:2.4
• Training service (local|remote|pai|aml|etc):
• Client OS:
• Server OS (for remote mode only):
• Python version:3.7
• PyTorch/TensorFlow version:PyTorch
• Is conda/virtualenv/venv used?:conda
• Is running in Docker?:

Configuration:

• Experiment config (remember to remove secrets!):
• Search space:

Log message:

• nnimanager.log:
• dispatcher.log:
• nnictl stdout and stderr:

/opt/conda/envs/pytorch1.9/lib/python3.7/site-packages/deprecate/deprecation.py:115: LightningDeprecationWarning: The Accuracy was deprecated since v1.3.0 in favor of torchmetrics.classification.accuracy.Accuracy. It will be removed in v1.5.0.
stream(template_mgs % msg_args)
Traceback (most recent call last):
File "/opt/conda/envs/pytorch1.9/lib/python3.7/runpy.py", line 193, in _run_module_as_main
"main", mod_spec)
File "/opt/conda/envs/pytorch1.9/lib/python3.7/runpy.py", line 85, in _run_code
exec(code, run_globals)
File "/home/notebook/code/personal/gzh/compression_toolkit/nni_v24/nni/nni/retiarii/trial_entry.py", line 25, in
engine.trial_execute_graph()
File "/home/notebook/code/personal/gzh/compression_toolkit/nni_v24/nni/nni/retiarii/execution/base.py", line 128, in trial_execute_graph
graph_data.evaluator._execute(model_cls)
File "/home/notebook/code/personal/gzh/compression_toolkit/nni_v24/nni/nni/retiarii/evaluator/pytorch/lightning.py", line 100, in _execute
return self.fit(model_cls)
File "/home/notebook/code/personal/gzh/compression_toolkit/nni_v24/nni/nni/retiarii/evaluator/pytorch/lightning.py", line 128, in fit
self.module.set_model(model)
File "/home/notebook/code/personal/gzh/compression_toolkit/nni_v24/nni/nni/retiarii/evaluator/pytorch/lightning.py", line 37, in set_model
self.model = model()
File "/home/notebook/code/personal/gzh/compression_toolkit/nni_v24/nni/examples/nas/oneshot/spos/_generated_model/G7UFOW.py", line 5580, in init
self.__first_conv = _model__first_conv(input_shape=[[1, 3, 32, 32]], output_shape=[[1, 16, 16, 16]])
TypeError: init() got an unexpected keyword argument 'input_shape'

我找到了G7UFOW.py文件，发现_model__first_conv类的构造函数没有入参（如下），确实有bug：

How to reproduce it?: just run

[ultmaster]

Delegate to nn-meter team for support.

[JiahangXu]

Delegate to nn-meter team for support.

Copy that. We will handle it as soon as possible.

[scarlett2018]

Delegate to nn-meter team for support.

Copy that. We will handle it as soon as possible.

@JiahangXu - any updates for this issue? thanks.

[JiahangXu]

Hi GeorgeW9527 and Scarlett, sorry for the late reply. This issue is in the process. The dev branch for nni is https://github.com/JiahangXu/nni/tree/dev-refine-shape-attr and for nn_meter is https://github.com/microsoft/nn-Meter/tree/dev/refine-shape-attr. The PR will be ready in one day.