[Bug]: offload/immediate_saving issue

[wenhuach21]

Problem Description

with pr #1656

/models/gemma-4-26B-A4B-it --layer_config '{"model.language_model.layers.\d+.self_attn..":{"bits":"8"},"model.language_model.layers.\d+.mlp..":{"bits":"8"},"model.language_model.layers.\d+.router..*":{"bits":"8"}}' --iters 0 --output_dir "/data2/wenhuach/mixed" --scheme W4A16

with --disable_opt_rtn is fine
/models/gemma-4-26B-A4B-it --layer_config '{"model.language_model.layers.\d+.self_attn..":{"bits":"8"},"model.language_model.layers.\d+.mlp..":{"bits":"8"},"model.language_model.layers.\d+.router..*":{"bits":"8"}}' --iters 0 --output_dir "/data2/wenhuach/mixed" --scheme W4A16 --disable_opt_rtn

Traceback (most recent call last):
File "/home/wenhuach/auto-round/auto_round/main.py", line 838, in
run()
~~~^^
File "/home/wenhuach/auto-round/auto_round/main.py", line 822, in run
start()
~~~~~^^
File "/home/wenhuach/auto-round/auto_round/main.py", line 541, in start
tune(args)
~~~~^^^^^^
File "/home/wenhuach/auto-round/auto_round/main.py", line 761, in tune
model, folders = autoround.quantize_and_save(export_dir, format=args.format) # pylint: disable=E1101
~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/wenhuach/auto-round/auto_round/compressors/base.py", line 1026, in quantize_and_save
model, _ = self.quantize()
~~~~~~~~~~~~~^^
File "/home/wenhuach/auto-round/auto_round/compressors/base.py", line 1816, in quantize
return self._quantize_rtn()
~~~~~~~~~~~~~~~~~~^^
File "/home/wenhuach/miniforge3/envs/autoround/lib/python3.13/site-packages/torch/utils/_contextlib.py", line 124, in decorate_context
return func(*args, **kwargs)
File "/home/wenhuach/auto-round/auto_round/compressors/base.py", line 1511, in _quantize_rtn
shard_writer(self, is_finalize=True)
~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/wenhuach/miniforge3/envs/autoround/lib/python3.13/site-packages/torch/utils/_contextlib.py", line 124, in decorate_context
return func(*args, **kwargs)
File "/home/wenhuach/auto-round/auto_round/compressors/shard_writer.py", line 267, in shard_writer
rounder._shard_writer.finalize()
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
File "/home/wenhuach/auto-round/auto_round/compressors/shard_writer.py", line 207, in finalize
self._add_tensor(pname, tensor.detach().to("cpu"))
~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/wenhuach/auto-round/auto_round/compressors/shard_writer.py", line 127, in _add_tensor
self._flush_shard()
~~~~~~~~~~~~~~~~~^^
File "/home/wenhuach/auto-round/auto_round/compressors/shard_writer.py", line 159, in _flush_shard
self._offload_to_meta(saved_params)
~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^
File "/home/wenhuach/auto-round/auto_round/compressors/shard_writer.py", line 176, in _offload_to_meta
module.to("meta")
~~~~~~~~~^^^^^^^^
File "/home/wenhuach/miniforge3/envs/autoround/lib/python3.13/site-packages/torch/nn/modules/module.py", line 1381, in to
return self._apply(convert)
~~~~~~~~~~~^^^^^^^^^
File "/home/wenhuach/miniforge3/envs/autoround/lib/python3.13/site-packages/torch/nn/modules/module.py", line 933, in _apply
module._apply(fn)
~~~~~~~~~~~~~^^^^
File "/home/wenhuach/miniforge3/envs/autoround/lib/python3.13/site-packages/torch/nn/modules/module.py", line 999, in _apply
assert isinstance(param, Parameter)
~~~~~~~~~~^^^^^^^^^^^^^^^^^^
AssertionError

Reproduction Steps

~

Environment Information

No response

Error Logs

Additional Context

No response

[lvliang-intel]

fixed in commit 4c087b6