Shortcuts

MinMaxObserver

class torch.ao.quantization.observer.MinMaxObserver(dtype=torch.quint8, qscheme=torch.per_tensor_affine, reduce_range=False, quant_min=None, quant_max=None, factory_kwargs=None, eps=1.1920928955078125e-07)[source]

Observer module for computing the quantization parameters based on the running min and max values.

This observer uses the tensor min/max statistics to compute the quantization parameters. The module records the running minimum and maximum of incoming tensors, and uses this statistic to compute the quantization parameters.

Parameters
  • dtype – dtype argument to the quantize node needed to implement the reference model spec.

  • qscheme – Quantization scheme to be used

  • reduce_range – Reduces the range of the quantized data type by 1 bit

  • quant_min – Minimum quantization value. If unspecified, it will follow the 8-bit setup.

  • quant_max – Maximum quantization value. If unspecified, it will follow the 8-bit setup.

  • eps (Tensor) – Epsilon value for float32, Defaults to torch.finfo(torch.float32).eps.

Given running min/max as xminx_\text{min} and xmaxx_\text{max}, scale ss and zero point zz are computed as:

The running minimum/maximum xmin/maxx_\text{min/max} is computed as:

xmin={min(X)if xmin=Nonemin(xmin,min(X))otherwisexmax={max(X)if xmax=Nonemax(xmax,max(X))otherwise\begin{array}{ll} x_\text{min} &= \begin{cases} \min(X) & \text{if~}x_\text{min} = \text{None} \\ \min\left(x_\text{min}, \min(X)\right) & \text{otherwise} \end{cases}\\ x_\text{max} &= \begin{cases} \max(X) & \text{if~}x_\text{max} = \text{None} \\ \max\left(x_\text{max}, \max(X)\right) & \text{otherwise} \end{cases}\\ \end{array}

where XX is the observed tensor.

The scale ss and zero point zz are then computed as:

if Symmetric:s=2max(xmin,xmax)/(QmaxQmin)z={0if dtype is qint8128otherwiseOtherwise:s=(xmaxxmin)/(QmaxQmin)z=Qminround(xmin/s)\begin{aligned} \text{if Symmetric:}&\\ &s = 2 \max(|x_\text{min}|, x_\text{max}) / \left( Q_\text{max} - Q_\text{min} \right) \\ &z = \begin{cases} 0 & \text{if dtype is qint8} \\ 128 & \text{otherwise} \end{cases}\\ \text{Otherwise:}&\\ &s = \left( x_\text{max} - x_\text{min} \right ) / \left( Q_\text{max} - Q_\text{min} \right ) \\ &z = Q_\text{min} - \text{round}(x_\text{min} / s) \end{aligned}

where QminQ_\text{min} and QmaxQ_\text{max} are the minimum and maximum of the quantized data type.

Warning

dtype can only take torch.qint8 or torch.quint8.

Note

If the running minimum equals to the running maximum, the scale and zero_point are set to 1.0 and 0.

calculate_qparams()[source]

Calculates the quantization parameters.

forward(x_orig)[source]

Records the running minimum and maximum of x.

reset_min_max_vals()[source]

Resets the min/max values.

Docs

Access comprehensive developer documentation for PyTorch

View Docs

Tutorials

Get in-depth tutorials for beginners and advanced developers

View Tutorials

Resources

Find development resources and get your questions answered

View Resources