lattice-inference 0.7.2

Pure Rust transformer inference engine — safetensors loading, SIMD matmul, BGE/Qwen3 embeddings
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23

//! Model module index for BERT, BitNet config, cross-encoder, Qwen, Qwen3.5, and Qwen3.5 config, with BERT/CrossEncoder/Qwen/Qwen3.5 type re-exports.
pub mod bert;
pub mod bitnet_config;
pub mod cross_encoder;
pub mod gemma4_cache;
pub mod gemma4_config;
mod gemma4_loading;
pub mod gemma4_model;
pub mod gemma4_ops;
pub mod gemma4_preflight;
mod gemma4_weights;
pub mod qwen;
pub mod qwen35;
pub mod qwen35_config;

// Re-export everything from bert (was top-level `model` module)
pub use self::bert::*;
pub use self::cross_encoder::CrossEncoderModel;
// Re-export key types from other model modules
pub use self::gemma4_model::Gemma4Model;
pub use self::qwen::{LayerTimings, ProfileTimings, QwenConfig, QwenModel};
pub use self::qwen35::{HiddenPooling, Qwen35Model};
pub use self::qwen35_config::GenerateConfig;