oxionnx 0.1.2

Pure Rust ONNX inference engine — zero C/C++ dependencies
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
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107

//! oxionnx — Pure Rust ONNX inference engine.
//!
//! Zero C/C++ dependencies. Supports a wide subset of ONNX operators
//! sufficient to run transformer-based models (BERT, T5, GPT-2 family).
//!
//! # Quick start
//!
//! ```no_run
//! use std::collections::HashMap;
//! use oxionnx::{Session, Tensor};
//!
//! fn main() -> Result<(), Box<dyn std::error::Error>> {
//!     let session = Session::from_file("model.onnx".as_ref())?;
//!     let mut inputs = HashMap::new();
//!     inputs.insert("input_ids", Tensor::new(vec![1.0, 2.0, 3.0], vec![1, 3]));
//!     let outputs = session.run(&inputs)?;
//!     Ok(())
//! }
//! ```

#[cfg(feature = "gpu")]
pub mod gpu {
    pub use oxionnx_gpu::*;
}
#[cfg(feature = "cuda")]
pub mod cuda {
    //! CUDA-accelerated dispatch for ONNX ops.
    pub use oxionnx_cuda::CudaContext;
    pub use oxionnx_cuda::CudaError;
}
#[cfg(feature = "cuda")]
pub use oxionnx_cuda::CudaContext;
#[cfg(feature = "directml")]
pub mod directml {
    //! DirectML (Windows D3D12 GPU) dispatch for ONNX ops.
    pub use oxionnx_directml::*;
}
pub mod graph {
    pub use oxionnx_core::graph::*;
}
pub mod model {
    pub use oxionnx_proto::model::*;
}
pub mod benchmark_select;
pub mod compat;
pub mod execution_providers;
pub mod io_binding;
pub mod macros;
pub mod memory;
pub mod memory_tracker;
#[cfg(feature = "ndarray")]
pub mod ndarray_compat;
pub mod optimizer;
pub mod ops {
    pub use oxionnx_ops::*;
}
pub mod proto {
    pub use oxionnx_proto::parser::*;
    pub use oxionnx_proto::streaming_parser::*;
    pub use oxionnx_proto::types::*;
}
#[cfg(feature = "encryption")]
pub mod encryption;
#[cfg(feature = "wasm")]
pub mod wasm;
#[cfg(feature = "mmap")]
pub mod mmap {
    pub use oxionnx_proto::mmap_loader::*;
}
pub mod opset_coverage;
pub mod session;
pub mod tolerance;
pub mod typed_session;
pub mod tensor {
    pub use oxionnx_core::tensor::*;
}

pub use benchmark_select::{
    benchmark_elementwise, benchmark_matmul, benchmark_suite, format_results, BenchmarkResult,
    ExecutionPath,
};
pub use compat::{GraphOptimizationLevel, SessionOutputs};
pub use execution_providers::{
    CPUExecutionProvider, CUDAExecutionProvider, CoreMLExecutionProvider,
    DirectMLExecutionProvider, ExecutionProviderDispatch, OpenVINOExecutionProvider,
    TensorRTExecutionProvider,
};
pub use io_binding::IoBinding;
pub use memory::{BufferPool, MemoryPlan, PoolStats, SizeClassPool};
pub use memory_tracker::MemoryTracker;
pub use oxionnx_core::OnnxError;
pub use oxionnx_core::OnnxError as Error;
pub use oxionnx_core::Tensor;
pub use oxionnx_core::{DType, TensorInfo, TensorStorage, TypedTensor};
#[cfg(feature = "gpu")]
pub use session::GpuExecutionProvider;
pub use session::{ModelInfo, ModelMetadata, NodeProfile, OptLevel, Session, SessionBuilder};
pub use tolerance::{compare_tensors, tensors_close, ToleranceReport};
pub use typed_session::{Shape, TypedSession};

// Re-export core types for convenience
pub use oxionnx_core::{
    Attributes, Dim, Graph, Node, OpContext, OpKind, Operator, OperatorRegistry,
};

// Re-export registry factory
pub use oxionnx_ops::default_registry;