Perform clip by value on the quantized Tensor operand
.
tf.raw_ops.UniformQuantizedClipByValue(
operand,
min,
max,
scales,
zero_points,
quantization_min_val,
quantization_max_val,
quantization_axis=-1,
name=None
)
Given quantized operand
which was quantized using scales
and zero_points
, performs clip by value using min
and max
values.
If quantization_axis is -1 (per-tensor quantized), the entire operand is clipped using scalar min, max.
Otherwise (per-channel quantized), the clipping is also done per-channel.
Args |
operand
|
A Tensor . Must be one of the following types: qint32 .
Must be a Tensor of T.
|
min
|
A Tensor . Must have the same type as operand .
The min value(s) to clip operand. Must be a Tensor of T.
Must be a scalar Tensor if quantization_axis is -1 (per-tensor quantization), otherwise 1D Tensor of size (operand.dim_size(quantization_axis),) (per-axis quantization).
|
max
|
A Tensor . Must have the same type as operand .
The min value(s) to clip operand. Must be a Tensor of T.
Must be a scalar Tensor if quantization_axis is -1 (per-tensor quantization), otherwise 1D Tensor of size (operand.dim_size(quantization_axis),) (per-axis quantization).
|
scales
|
A Tensor of type float32 .
The float value(s) used as scale(s) when quantizing operand , min and max .
Must be a scalar Tensor if quantization_axis is -1 (per-tensor quantization), otherwise 1D Tensor of size (operand.dim_size(quantization_axis),) (per-axis quantization).
|
zero_points
|
A Tensor of type int32 .
The int32 value(s) used as zero_point(s) when quantizing operand , min and max .
Same shape condition as scales.
|
quantization_min_val
|
An int .
The quantization min value that was used when operand was quantized.
|
quantization_max_val
|
An int .
The quantization max value that was used when operand was quantized.
|
quantization_axis
|
An optional int . Defaults to -1 .
Indicates the dimension index of the tensor where per-axis quantization is applied for the slices along that dimension.
If set to -1 (default), this indicates per-tensor quantization. Otherwise, it must be set within range [0, operand.dims()).
|
name
|
A name for the operation (optional).
|
Returns |
A Tensor . Has the same type as operand .
|