wgml 0.2.0

Cross-platform GPU LLM inference.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19

//! Primitives for building LLM inferences.

mod batched_multiquery_attention;
mod layernorm;
mod rms_norm;
mod rope;
mod silu;
mod softmax;
mod unary;

pub use batched_multiquery_attention::{
    BatchedMultiqueryAttention, BatchedMultiqueryAttentionParams,
};
pub use layernorm::LayerNorm;
pub use rms_norm::RmsNorm;
pub use rope::{RoPE, RoPEShape};
pub use silu::Silu;
pub use softmax::SoftMax;
pub use unary::{Unary, UnaryOp};