【寒武纪硬件产品型号】必填*：MLU220

【使用操作系统】必填*：Ubuntu 18.04.5 LTS 

【使用驱动版本】必填*：CNRT: 4.10.2 adb709e

【问题】：经过pytorch量化感知训练后的模型，支持转换成MLU220离线模型吗？

【错误信息】：

Traceback (most recent call last):

  File "mlu_forward_qat.py", line 347, in <module>

    mlu_forward( opt.img, use_mlu=True, mlu220 = save_offline_model, fusion= opt.fusion)

  File "mlu_forward_qat.py", line 237, in mlu_forward

    model = mlu_quantize.quantize_dynamic_mlu(model)

  File "/torch/venv3/pytorch/lib/python3.6/site-packages/torch_mlu/core/mlu_quantize.py", line 236, in quantize_dynamic_mlu

    model = copy.deepcopy(model)

  File "/usr/local/python3.6/lib/python3.6/copy.py", line 180, in deepcopy

    y = _reconstruct(x, memo, *rv)

  File "/usr/local/python3.6/lib/python3.6/copy.py", line 280, in _reconstruct

    state = deepcopy(state, memo)

  File "/usr/local/python3.6/lib/python3.6/copy.py", line 150, in deepcopy

    y = copier(x, memo)

  File "/usr/local/python3.6/lib/python3.6/copy.py", line 240, in _deepcopy_dict

    y[deepcopy(key, memo)] = deepcopy(value, memo)

  File "/usr/local/python3.6/lib/python3.6/copy.py", line 180, in deepcopy

    y = _reconstruct(x, memo, *rv)

  File "/usr/local/python3.6/lib/python3.6/copy.py", line 306, in _reconstruct

    value = deepcopy(value, memo)

  File "/usr/local/python3.6/lib/python3.6/copy.py", line 180, in deepcopy

    y = _reconstruct(x, memo, *rv)

  File "/usr/local/python3.6/lib/python3.6/copy.py", line 280, in _reconstruct

    state = deepcopy(state, memo)

  File "/usr/local/python3.6/lib/python3.6/copy.py", line 150, in deepcopy

    y = copier(x, memo)

  File "/usr/local/python3.6/lib/python3.6/copy.py", line 240, in _deepcopy_dict

    y[deepcopy(key, memo)] = deepcopy(value, memo)

  File "/usr/local/python3.6/lib/python3.6/copy.py", line 180, in deepcopy

    y = _reconstruct(x, memo, *rv)

  File "/usr/local/python3.6/lib/python3.6/copy.py", line 306, in _reconstruct

    value = deepcopy(value, memo)

  File "/usr/local/python3.6/lib/python3.6/copy.py", line 180, in deepcopy

    y = _reconstruct(x, memo, *rv)

  File "/usr/local/python3.6/lib/python3.6/copy.py", line 280, in _reconstruct

    state = deepcopy(state, memo)

  File "/usr/local/python3.6/lib/python3.6/copy.py", line 150, in deepcopy

    y = copier(x, memo)

  File "/usr/local/python3.6/lib/python3.6/copy.py", line 240, in _deepcopy_dict

    y[deepcopy(key, memo)] = deepcopy(value, memo)

  File "/usr/local/python3.6/lib/python3.6/copy.py", line 180, in deepcopy

    y = _reconstruct(x, memo, *rv)

  File "/usr/local/python3.6/lib/python3.6/copy.py", line 306, in _reconstruct

    value = deepcopy(value, memo)

  File "/usr/local/python3.6/lib/python3.6/copy.py", line 169, in deepcopy

    rv = reductor(4)

  File "/torch/venv3/pytorch/lib/python3.6/site-packages/torch/nn/quantized/modules/conv.py", line 151, in __getstate__

    raise RuntimeError('torch.save() is not currently supported for quantized modules.'

RuntimeError: torch.save() is not currently supported for quantized modules. See https://github.com/pytorch/pytorch/issues/24045. Please use state_dict or torch.jit serialization.