Skip to main content

dequantize_weights

Function dequantize_weights 

Source
pub fn dequantize_weights(
    ternary: &TernaryWeight,
    device: &Device,
) -> Result<Tensor>
Expand description

Dequantize ternary weights back to float tensor.

§Arguments

  • ternary - Ternary weight to dequantize
  • device - Device to create output tensor on

§Errors

Returns error if tensor creation fails.