Function quantize_f16

Source
pub fn quantize_f16(checkpoint: &mut HashMap<String, Tensor>) -> Result<()>
Expand description

Quantize the float tensors to f16.