Cast#

Cast - 13 #

Version

name: Cast (GitHub)
domain: main
since_version: 13
function: False
support_level: SupportType.COMMON
shape inference: True

This version of the operator has been available since version 13.

Summary

The operator casts the elements of a given input tensor to a data type specified by the ‘to’ argument and returns an output tensor of the same size in the converted type. The ‘to’ argument must be one of the data types specified in the ‘DataType’ enum field in the TensorProto message.

Casting from string tensor in plain (e.g., “3.14” and “1000”) and scientific numeric representations (e.g., “1e-5” and “1E8”) to float types is supported. For example, converting string “100.5” to an integer may result 100. There are some string literals reserved for special floating-point values; “+INF” (and “INF”), “-INF”, and “NaN” are positive infinity, negative infinity, and not-a-number, respectively. Any string which can exactly match “+INF” in a case-insensitive way would be mapped to positive infinite. Similarly, this case-insensitive rule is applied to “INF” and “NaN”. When casting from numeric tensors to string tensors, plain floating-point representation (such as “314.15926”) would be used. Converting non-numerical-literal string such as “Hello World!” is an undefined behavior. Cases of converting string representing floating-point arithmetic value, such as “2.718”, to INT is an undefined behavior.

Conversion from a numerical type to any numerical type is always allowed. User must be aware of precision loss and value change caused by range difference between two types. For example, a 64-bit float 3.1415926459 may be round to a 32-bit float 3.141592. Similarly, converting an integer 36 to Boolean may produce 1 because we truncate bits which can’t be stored in the targeted type.

In more detail, the conversion among numerical types should follow these rules:

Casting from floating point to: * floating point: +/- infinity if OOR (out of range). * fixed point: undefined if OOR. * bool: +/- 0.0 to False; all else to True.
Casting from fixed point to: * floating point: +/- infinity if OOR. (+ infinity in the case of uint) * fixed point: when OOR, discard higher bits and reinterpret (with respect to two’s complement representation for

signed types). For example, 200 (int16) -> -56 (int8).

bool: zero to False; nonzero to True.

Casting from bool to: * floating point: {1.0, 0.0}. * fixed point: {1, 0}. * bool: no change.

Attributes

to (required): The data type to which the elements of the input tensor are cast. Strictly must be one of the types from DataType enum in TensorProto

Inputs

input (heterogeneous) - T1: Input tensor to be cast.

Outputs

output (heterogeneous) - T2: Output tensor with the same shape as input with type specified by the ‘to’ argument

Type Constraints

T1 in ( tensor(bfloat16), tensor(bool), tensor(double), tensor(float), tensor(float16), tensor(int16), tensor(int32), tensor(int64), tensor(int8), tensor(string), tensor(uint16), tensor(uint32), tensor(uint64), tensor(uint8) ): Constrain input types. Casting from complex is not supported.
T2 in ( tensor(bfloat16), tensor(bool), tensor(double), tensor(float), tensor(float16), tensor(int16), tensor(int32), tensor(int64), tensor(int8), tensor(string), tensor(uint16), tensor(uint32), tensor(uint64), tensor(uint8) ): Constrain output types. Casting to complex is not supported.

Examples

default

import numpy as np
import onnx

shape = (3, 4)
test_cases = [
    ("FLOAT", "FLOAT16"),
    ("FLOAT", "DOUBLE"),
    ("FLOAT16", "FLOAT"),
    ("FLOAT16", "DOUBLE"),
    ("DOUBLE", "FLOAT"),
    ("DOUBLE", "FLOAT16"),
    ("FLOAT", "STRING"),
    ("STRING", "FLOAT"),
    ("FLOAT", "BFLOAT16"),
    ("BFLOAT16", "FLOAT"),
]

for from_type, to_type in test_cases:
    input_type_proto = None
    output_type_proto = None
    if "BFLOAT16" == from_type or "BFLOAT16" == to_type:
        np_fp32 = np.array(
            [
                "0.47892547",
                "0.48033667",
                "0.49968487",
                "0.81910545",
                "0.47031248",
                "0.816468",
                "0.21087195",
                "0.7229038",
                "NaN",
                "INF",
                "+INF",
                "-INF",
            ],
            dtype=np.float32,
        )
        little_endisan = sys.byteorder == "little"
        np_uint16_view = np_fp32.view(dtype=np.uint16)
        np_bfp16 = (
            np_uint16_view[1::2] if little_endisan else np_uint16_view[0::2]
        )
        if "BFLOAT16" == to_type:
            assert from_type == "FLOAT"
            input = np_fp32.reshape([3, 4])
            output = np_bfp16.reshape([3, 4])
            input_type_proto = onnx.helper.make_tensor_type_proto(
                int(TensorProto.FLOAT), input.shape
            )
            output_type_proto = onnx.helper.make_tensor_type_proto(
                int(TensorProto.BFLOAT16), output.shape
            )
        else:
            assert to_type == "FLOAT"
            input = np_bfp16.reshape([3, 4])
            # convert bfloat to FLOAT
            np_fp32_zeros = np.zeros((len(np_bfp16) * 2,), dtype=np.uint16)
            if little_endisan:
                np_fp32_zeros[1::2] = np_bfp16
            else:
                np_fp32_zeros[0::2] = np_bfp16
            np_fp32_from_bfloat = np_fp32_zeros.view(dtype=np.float32)
            output = np_fp32_from_bfloat.reshape([3, 4])
            input_type_proto = onnx.helper.make_tensor_type_proto(
                int(TensorProto.BFLOAT16), input.shape
            )
            output_type_proto = onnx.helper.make_tensor_type_proto(
                int(TensorProto.FLOAT), output.shape
            )
    elif "STRING" != from_type:
        input = np.random.random_sample(shape).astype(
            helper.tensor_dtype_to_np_dtype(getattr(TensorProto, from_type))
        )
        if "STRING" == to_type:
            # Converting input to str, then give it object dtype for generating script
            ss = []
            for i in input.flatten():
                s = str(i).encode("utf-8")
                su = s.decode("utf-8")
                ss.append(su)

            output = np.array(ss).astype(object).reshape([3, 4])
        else:
            output = input.astype(
                helper.tensor_dtype_to_np_dtype(getattr(TensorProto, to_type))
            )
    else:
        input = np.array(
            [
                "0.47892547",
                "0.48033667",
                "0.49968487",
                "0.81910545",
                "0.47031248",
                "0.816468",
                "0.21087195",
                "0.7229038",
                "NaN",
                "INF",
                "+INF",
                "-INF",
            ],
            dtype=np.dtype(object),
        ).reshape([3, 4])
        output = input.astype(
            helper.tensor_dtype_to_np_dtype(getattr(TensorProto, to_type))
        )
    node = onnx.helper.make_node(
        "Cast",
        inputs=["input"],
        outputs=["output"],
        to=getattr(TensorProto, to_type),
    )
    if input_type_proto and output_type_proto:
        expect(
            node,
            inputs=[input],
            outputs=[output],
            name="test_cast_" + from_type + "_to_" + to_type,
            input_type_protos=[input_type_proto],
            output_type_protos=[output_type_proto],
        )
    else:
        expect(
            node,
            inputs=[input],
            outputs=[output],
            name="test_cast_" + from_type + "_to_" + to_type,
        )

Cast - 9 vs 13

Cast - 9 #

Version

name: Cast (GitHub)
domain: main
since_version: 9
function: False
support_level: SupportType.COMMON
shape inference: True

This version of the operator has been available since version 9.

Summary

Attributes

to (required): The data type to which the elements of the input tensor are cast. Strictly must be one of the types from DataType enum in TensorProto

Inputs

input (heterogeneous) - T1: Input tensor to be cast.

Outputs

output (heterogeneous) - T2: Output tensor with the same shape as input with type specified by the ‘to’ argument

Type Constraints

T1 in ( tensor(bool), tensor(double), tensor(float), tensor(float16), tensor(int16), tensor(int32), tensor(int64), tensor(int8), tensor(string), tensor(uint16), tensor(uint32), tensor(uint64), tensor(uint8) ): Constrain input types. Casting from complex is not supported.
T2 in ( tensor(bool), tensor(double), tensor(float), tensor(float16), tensor(int16), tensor(int32), tensor(int64), tensor(int8), tensor(string), tensor(uint16), tensor(uint32), tensor(uint64), tensor(uint8) ): Constrain output types. Casting to complex is not supported.

Cast - 6 #

Version

name: Cast (GitHub)
domain: main
since_version: 6
function: False
support_level: SupportType.COMMON
shape inference: True

This version of the operator has been available since version 6.

Summary

Attributes

to (required): The data type to which the elements of the input tensor are cast. Strictly must be one of the types from DataType enum in TensorProto

Inputs

input (heterogeneous) - T1: Input tensor to be cast.

Outputs

output (heterogeneous) - T2: Output tensor with the same shape as input with type specified by the ‘to’ argument

Type Constraints

T1 in ( tensor(bool), tensor(double), tensor(float), tensor(float16), tensor(int16), tensor(int32), tensor(int64), tensor(int8), tensor(uint16), tensor(uint32), tensor(uint64), tensor(uint8) ): Constrain input types. Casting from strings and complex are not supported.
T2 in ( tensor(bool), tensor(double), tensor(float), tensor(float16), tensor(int16), tensor(int32), tensor(int64), tensor(int8), tensor(uint16), tensor(uint32), tensor(uint64), tensor(uint8) ): Constrain output types. Casting to strings and complex are not supported.

Cast - 1 #

Version

name: Cast (GitHub)
domain: main
since_version: 1
function: False
support_level: SupportType.COMMON
shape inference: False

This version of the operator has been available since version 1.

Summary

Attributes

to (required): The data type to which the elements of the input tensor are cast. Strictly must be one of the types from DataType enum in TensorProto

Inputs

input (heterogeneous) - T1: Input tensor to be cast.

Outputs

output (heterogeneous) - T2: Output tensor with the same shape as input with type specified by the ‘to’ argument

Type Constraints

T1 in ( tensor(bool), tensor(double), tensor(float), tensor(float16), tensor(int16), tensor(int32), tensor(int64), tensor(int8), tensor(uint16), tensor(uint32), tensor(uint64), tensor(uint8) ): Constrain input types. Casting from strings and complex are not supported.
T2 in ( tensor(bool), tensor(double), tensor(float), tensor(float16), tensor(int16), tensor(int32), tensor(int64), tensor(int8), tensor(uint16), tensor(uint32), tensor(uint64), tensor(uint8) ): Constrain output types. Casting to strings and complex are not supported.

Cast#

Cast - 13#

Cast - 9#

Cast - 6#

Cast - 1#

Cast - 13 #

Cast - 9 #

Cast - 6 #

Cast - 1 #