Enum Documentation¶

enum paddle::ScaleAlgo¶

Algorithms for finding scale of quantized Tensors.

Values:

enumerator NONE¶

Do not compute scale.

enumerator MAX¶

Find scale based on the max absolute value.

enumerator MAX_CH¶

Find scale based on the max absolute value per output channel.

enumerator MAX_CH_T¶

Find scale based on the max absolute value per output channel of a transposed tensor

enumerator KL¶

Find scale based on KL Divergence.