quantize

aimet_torch.quantization.affine.quantize(tensor, scale, offset, *args, **kwargs)

Applies quantization to the input.

Precisely,

$$out=clamp(\lceil \frac{input}{scale}\rfloor -offset,qmin,qmax)$$

If block size $B=\left(\begin{array}{cccc}{B}_0& B_1& \cdots & B_{D-1}\end{array}\right)$ is specified, this equation will be further generalized as

$$\begin{array}{r}\begin{array}{c}\begin{array}{rl}ou{t}_{j_0\cdots j_{D-1}}& =clamp(\lceil \frac{inpu{t}_{j_0\cdots j_{D-1}}}{scal{e}_{i_0\cdots i_{D-1}}}\rfloor -offse{t}_{i_0\cdots i_{D-1}},qmin,qmax)\end{array}\\ \text{where}{\mathrm{\forall }}_{0\le d<D}{i}_d=\lfloor \frac{j_d}{B_d}\rfloor \end{array}\end{array}$$

This function is overloaded with the signatures listed below:

aimet_torch.quantization.affine.quantize(tensor, scale, offset, bitwidth, signed=False, block_size=None)

Equivalent to:

$$\begin{array}{r}qmin=\{\begin{array}{ll}-\lceil \frac{2^{bitwidth}-1}{2}\rceil ,& \text{if }signed\\ 0,& \text{otherwise (default)}\end{array}qmax=\{\begin{array}{ll}\lfloor \frac{2^{bitwidth}-1}{2}\rfloor ,& \text{if }signed\\ 2^{bitwidth}-1,& \text{otherwise (default)}\end{array}\end{array}$$

Parameters:

• tensor (Tensor) – Tensor to quantize

• scale (Tensor) – Scale for quantization

• offset (Tensor) – Offset for quantization

• bitwidth (int) – Bitwidth of quantized tensor based on which $qmin$ and $qmax$ will be derived

• signed (bool) – If false, the output will be mapped to positive integers only. Otherwise, it will range over both positive and negative integers.

• block_size (Tuple[int, ...], optional) – Block size

aimet_torch.quantization.affine.quantize(tensor, scale, offset, *, num_steps, signed=False, block_size=None)

Equivalent to:

$$\begin{array}{r}qmin=\{\begin{array}{ll}-\lceil \frac{num\mathrm{\_}steps}{2}\rceil ,& \text{if }signed\\ 0,& \text{otherwise (default)}\end{array}qmax=\{\begin{array}{ll}\lfloor \frac{num\mathrm{\_}steps}{2}\rfloor ,& \text{if }signed\\ num\mathrm{\_}steps,& \text{otherwise (default)}\end{array}\end{array}$$

Parameters:

• tensor (Tensor) – Tensor to quantize

• scale (Tensor) – Scale for quantization

• offset (Tensor) – Offset for quantization

• num_steps (int) – The number of steps in the quantization range based on which $qmin$ and $qmax$ will be derived

• signed (bool) – If false, the output will be mapped to positive integers only. Otherwise, it will range over both positive and negative integers.

• block_size (Tuple[int, ...], optional) – Block size

aimet_torch.quantization.affine.quantize(tensor, scale, offset, *, qmin, qmax, block_size=None)

Parameters:

• tensor (Tensor) – Tensor to quantize

• scale (Tensor) – Scale for quantization

• offset (Tensor) – Offset for quantization

• qmin (int) – Minimum value of the quantization range

• qmax (int) – Maximum value of the quantization range

• block_size (Tuple[int, ...], optional) – Block size

Examples

>>> import aimet_torch.v2.quantization as Q
>>> input = torch.arange(start=-0.3, end=1.3, step=0.05)
>>> print(input)
tensor([-3.0000e-01, -2.5000e-01, -2.0000e-01, -1.5000e-01, -1.0000e-01,
        -5.0000e-02, -1.1921e-08,  5.0000e-02,  1.0000e-01,  1.5000e-01,
        2.0000e-01,  2.5000e-01,  3.0000e-01,  3.5000e-01,  4.0000e-01,
        4.5000e-01,  5.0000e-01,  5.5000e-01,  6.0000e-01,  6.5000e-01,
        7.0000e-01,  7.5000e-01,  8.0000e-01,  8.5000e-01,  9.0000e-01,
        9.5000e-01,  1.0000e+00,  1.0500e+00,  1.1000e+00,  1.1500e+00,
        1.2000e+00,  1.2500e+00])
>>> scale = torch.tensor(1/15)
>>> offset = torch.tensor(0.0)
>>> Q.affine.quantize(input, scale, offset, bitwidth=4)
tensor([ 0.,  0.,  0.,  0.,  0.,  0., -0.,  1.,  2.,  2.,  3.,  4.,  4.,  5.,
         6.,  7.,  7.,  8.,  9., 10., 10., 11., 12., 13., 13., 14., 15., 15.,
         15., 15., 15., 15.])
>>> Q.affine.quantize(input, scale, offset, num_steps=15)
tensor([ 0.,  0.,  0.,  0.,  0.,  0., -0.,  1.,  2.,  2.,  3.,  4.,  4.,  5.,
         6.,  7.,  7.,  8.,  9., 10., 10., 11., 12., 13., 13., 14., 15., 15.,
         15., 15., 15., 15.])
>>> Q.affine.quantize(input, scale, offset, qmin=0, qmax=15)
tensor([ 0.,  0.,  0.,  0.,  0.,  0., -0.,  1.,  2.,  2.,  3.,  4.,  4.,  5.,
         6.,  7.,  7.,  8.,  9., 10., 10., 11., 12., 13., 13., 14., 15., 15.,
         15., 15., 15., 15.])