rullama 0.4.0

Browser-resident Gemma 4 inference: pure Rust → WebAssembly + WebGPU. Loads Ollama's on-disk GGUF blobs and runs the forward pass on the local GPU via hand-written WGSL.
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
25
26
27
28
29
30
31
32
33

use thiserror::Error;

#[derive(Debug, Error)]
pub enum RullamaError {
    #[error("WebGPU is not available in this environment")]
    WebGpuUnavailable,

    #[error("failed to request a wgpu adapter")]
    NoAdapter,

    #[error("failed to request a wgpu device: {0}")]
    DeviceRequest(String),

    #[error("GGUF parse error: {0}")]
    Gguf(String),

    #[error("model config error: {0}")]
    Config(String),

    #[error("tokenizer error: {0}")]
    Tokenizer(String),

    #[error("inference error: {0}")]
    Inference(String),

    #[error("buffer mapping failed: {0}")]
    BufferMap(String),

    #[error("cancelled by caller")]
    Cancelled,
}

pub type Result<T, E = RullamaError> = core::result::Result<T, E>;