module 'numpy' has no attribute 'asscalar'

[zhangluustb]

File "/usr/local/lib/python3.8/dist-packages/ppq-0.6.4-py3.8.egg/ppq/parser/util.py", line 27, in convert_value
value = np.asscalar(value[0])
File "/usr/local/lib/python3.8/dist-packages/numpy/init.py", line 311, in getattr
raise AttributeError("module {!r} has no attribute "
AttributeError: module 'numpy' has no attribute 'asscalar'

numpy version 1.23.0

[ZhangZhiPku]

听说是numpy新版中移除了这个函数：

numpy.asscalar(a)
将大小为1的数组转换为其标量等效值。

1.16 版后已移除: 弃用，使用 numpy.ndarray.item() 相反。

[ZhangZhiPku]

问题将在下一次更新中修复，你可以使用下面的写法替换 util.py

def convert_value(
    value: Union[int, float, np.ndarray, torch.Tensor],
    export_as_float: bool, dtype: DataType = DataType.FP32) -> Union[float, list]:
    """Converting value from any to python native data dtype, ready for export.

    Args:
        value (Union[int, float, np.ndarray, torch.Tensor]): exporting value.
        export_as_list (bool): export value as a list.
        dtype (DataType, optional): exporting dtype.

    Returns:
        Union[float, list]: Converted value
    """
    if dtype not in {DataType.FP32, DataType.INT32}:
        raise ValueError(f'Can Only export dtype fp32 and int32, '
                         f'while you are requiring to dump a {dtype.name} value')
    value = convert_any_to_numpy(value, accepet_none=False)
    value = value.astype(dtype=DataType.to_numpy(dtype))
    if export_as_float:
        value = value[0].item()
        assert type(value) in {int, float}, (
            f'Trying to dump a tensorwise quantization value {value}. '
            f'It is Expected to be a int or float value, while {type(value)} was given')
        return value
    else:
        value = convert_any_to_numpy(value, accepet_none=False)
        return value.tolist()

[ZhangZhiPku]

问题已经在0.6.5的更新中修复