ML::GGUF::ComputeBackend
Abstract compute backend. All math ops go through this interface.
Instance methods
Dot product for attention scores
layer_norm!(x : Array(Float32), n_pos : Int32, dim : Int32, w : Array(Float32), b : Array(Float32)) : Nil
In-place layer norm
Matrix multiply: x[rows, in_dim] × W_quant + bias → [rows, out_dim]