qonnx.core.datatype (module)

class qonnx.core.datatype.BaseDataType

Bases: abc.ABC

Base class for QONNX data types.

abstract allowed(value)

Check whether given value is allowed for this DataType.

value (float32): value to be checked

abstract bitwidth(): Returns the number of bits required for this DataType.

abstract get_canonical_name(): Return a canonical string representation of this QONNX DataType.

abstract get_hls_datatype_str(): Returns the corresponding Vivado HLS datatype name.

abstract get_num_possible_values(): Returns the number of possible values this DataType can take. Only implemented for integer types for now.

abstract is_fixed_point(): Returns whether this DataType represent fixed-point values only.

abstract is_integer(): Returns whether this DataType represents integer values only.

abstract max(): Returns the largest possible value allowed by this DataType.

abstract min(): Returns the smallest possible value allowed by this DataType.

property name

signed(): Returns whether this DataType can represent negative numbers.

abstract to_numpy_dt(): Return an appropriate numpy datatype that can represent this QONNX DataType.

class qonnx.core.datatype.BipolarType

Bases: qonnx.core.datatype.BaseDataType

allowed(value)

Check whether given value is allowed for this DataType.

value (float32): value to be checked

bitwidth(): Returns the number of bits required for this DataType.

get_canonical_name(): Return a canonical string representation of this QONNX DataType.

get_hls_datatype_str(): Returns the corresponding Vivado HLS datatype name.

get_num_possible_values(): Returns the number of possible values this DataType can take. Only implemented for integer types for now.

is_fixed_point(): Returns whether this DataType represent fixed-point values only.

is_integer(): Returns whether this DataType represents integer values only.

max(): Returns the largest possible value allowed by this DataType.

min(): Returns the smallest possible value allowed by this DataType.

to_numpy_dt(): Return an appropriate numpy datatype that can represent this QONNX DataType.

class qonnx.core.datatype.DataType(value)

Bases: enum.Enum

Enum class that contains QONNX data types to set the quantization annotation. ONNX does not support data types smaller than 8-bit integers, whereas in QONNX we are interested in smaller integers down to ternary and bipolar.

static get_accumulator_dt_cands()

static get_smallest_possible(value): Returns smallest (fewest bits) possible DataType that can represent value. Prefers unsigned integers where possible.

class qonnx.core.datatype.DataTypeMeta(cls, bases, classdict): Bases: enum.EnumMeta

class qonnx.core.datatype.FixedPointType(bitwidth, intwidth)

Bases: qonnx.core.datatype.IntType

allowed(value)

Check whether given value is allowed for this DataType.

value (float32): value to be checked

frac_bits()

get_canonical_name(): Return a canonical string representation of this QONNX DataType.

get_hls_datatype_str(): Returns the corresponding Vivado HLS datatype name.

int_bits()

is_fixed_point(): Returns whether this DataType represent fixed-point values only.

is_integer(): Returns whether this DataType represents integer values only.

max(): Returns the largest possible value allowed by this DataType.

min(): Returns the smallest possible value allowed by this DataType.

scale_factor()

to_numpy_dt(): Return an appropriate numpy datatype that can represent this QONNX DataType.

class qonnx.core.datatype.FloatType

Bases: qonnx.core.datatype.BaseDataType

allowed(value)

Check whether given value is allowed for this DataType.

value (float32): value to be checked

bitwidth(): Returns the number of bits required for this DataType.

get_canonical_name(): Return a canonical string representation of this QONNX DataType.

get_hls_datatype_str(): Returns the corresponding Vivado HLS datatype name.

get_num_possible_values(): Returns the number of possible values this DataType can take. Only implemented for integer types for now.

is_fixed_point(): Returns whether this DataType represent fixed-point values only.

is_integer(): Returns whether this DataType represents integer values only.

max(): Returns the largest possible value allowed by this DataType.

min(): Returns the smallest possible value allowed by this DataType.

to_numpy_dt(): Return an appropriate numpy datatype that can represent this QONNX DataType.

class qonnx.core.datatype.IntType(bitwidth, signed)

Bases: qonnx.core.datatype.BaseDataType

allowed(value)

Check whether given value is allowed for this DataType.

value (float32): value to be checked

bitwidth(): Returns the number of bits required for this DataType.

get_canonical_name(): Return a canonical string representation of this QONNX DataType.

get_hls_datatype_str(): Returns the corresponding Vivado HLS datatype name.

get_num_possible_values(): Returns the number of possible values this DataType can take. Only implemented for integer types for now.

is_fixed_point(): Returns whether this DataType represent fixed-point values only.

is_integer(): Returns whether this DataType represents integer values only.

max(): Returns the largest possible value allowed by this DataType.

min(): Returns the smallest possible value allowed by this DataType.

to_numpy_dt(): Return an appropriate numpy datatype that can represent this QONNX DataType.

class qonnx.core.datatype.TernaryType

Bases: qonnx.core.datatype.BaseDataType

allowed(value)

Check whether given value is allowed for this DataType.

value (float32): value to be checked

bitwidth(): Returns the number of bits required for this DataType.

get_canonical_name(): Return a canonical string representation of this QONNX DataType.

get_hls_datatype_str(): Returns the corresponding Vivado HLS datatype name.

get_num_possible_values(): Returns the number of possible values this DataType can take. Only implemented for integer types for now.

is_fixed_point(): Returns whether this DataType represent fixed-point values only.

is_integer(): Returns whether this DataType represents integer values only.

max(): Returns the largest possible value allowed by this DataType.

min(): Returns the smallest possible value allowed by this DataType.

to_numpy_dt(): Return an appropriate numpy datatype that can represent this QONNX DataType.

qonnx.core.datatype.resolve_datatype(name)