forge-ml 0.1.0

A WebGPU-native machine learning runtime in Rust: train and run GPT-2 on any GPU wgpu reaches, with no CUDA and no Python
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24

// Elementwise addition: c = a + b (same shape).

struct Params {
    n: u32,
    _pad0: u32,
    _pad1: u32,
    _pad2: u32,
}

@group(0) @binding(0) var<uniform> p: Params;
@group(0) @binding(1) var<storage, read> a: array<f32>;
@group(0) @binding(2) var<storage, read> b: array<f32>;
@group(0) @binding(3) var<storage, read_write> c: array<f32>;

@compute @workgroup_size(256)
fn main(
    @builtin(workgroup_id) wid: vec3<u32>,
    @builtin(num_workgroups) nwg: vec3<u32>,
    @builtin(local_invocation_index) li: u32,
) {
    let i = (wid.y * nwg.x + wid.x) * 256u + li;
    if (i >= p.n) { return; }
    c[i] = a[i] + b[i];
}