Skip to content

reshape

Reshape

Bases: TensorOp

Reshape a input tensor to conform to a given shape.

Parameters:

Name Type Description Default
inputs Union[str, List[str]]

Key of the input tensor that is to be reshaped.

 required
outputs Union[str, List[str]]

Key of the output tensor that has been reshaped.

 required
shape Union[int, Tuple[int, ...]]

Target shape.

 required
mode Union[None, str, Iterable[str]]

What mode(s) to execute this Op in. For example, "train", "eval", "test", or "infer". To execute regardless of mode, pass None. To execute in all modes except for a particular one, you can pass an argument like "!infer" or "!train".

 '!infer'
Source code in fastestimator\fastestimator\op\tensorop\reshape.py 
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
class Reshape(TensorOp):
    """Reshape a input tensor to conform to a given shape.

    Args:
        inputs: Key of the input tensor that is to be reshaped.
        outputs: Key of the output tensor that has been reshaped.
        shape: Target shape.
        mode: What mode(s) to execute this Op in. For example, "train", "eval", "test", or "infer". To execute
            regardless of mode, pass None. To execute in all modes except for a particular one, you can pass an argument
            like "!infer" or "!train".
    """
    def __init__(self,
                 inputs: Union[str, List[str]],
                 outputs: Union[str, List[str]],
                 shape: Union[int, Tuple[int, ...]],
                 mode: Union[None, str, Iterable[str]] = "!infer") -> None:

        super().__init__(inputs=inputs, outputs=outputs, mode=mode)
        self.shape = shape
        self.in_list, self.out_list = False, False

    def forward(self, data: Tensor, state: Dict[str, Any]) -> Tensor:
        if isinstance(data, tf.Tensor):
            return tf.reshape(data, self.shape)
        elif isinstance(data, torch.Tensor):
            return data.view(self.shape)
        else:
            raise ValueError("unrecognized data format: {}".format(type(data)))