struct

ML::GGUF::QuantWeight

Inherits Struct < Value < Object

Quantized weight: raw bytes + type for on-the-fly dequant during matmul

Constructors

new(raw : Slice(UInt8), type : ML::GGUF::TensorType, out_dim : Int32, in_dim : Int32)

Instance methods

in_dim

out_dim

raw

type