oxicuda-vision 0.2.0

Vision Transformer & CLIP primitives for OxiCUDA: ViT patch embedding, multi-head self-attention, CLIP contrastive learning, FPN, RoI align, DETR decoder — pure Rust, zero CUDA SDK dependency.
Documentation 
1
2
3
4
5
6
7
8
9
10

//! Point-cloud neural network primitives.
//!
//! Provides:
//! - **`point_transformer`**: the Point Transformer vector self-attention layer
//!   (Zhao et al. 2021) — subtraction-relation attention over kNN
//!   neighbourhoods with a learned relative-position encoding.

pub mod point_transformer;

pub use point_transformer::{PointAttention, PointTransformerConfig, PointTransformerLayer};