Rate this Page
★★★★★

torch.Tensor.q_per_channel_zero_points#

Tensor.q_per_channel_zero_points()→Tensor#

Given a Tensor quantized by linear (affine) per-channel quantization,returns a tensor of zero_points of the underlying quantizer. It has the number ofelements that matches the corresponding dimensions (from q_per_channel_axis) ofthe tensor.