--- title: "onemodel docs" url: https://recursiveai.org/onemodel/docs/ description: "Everything about OneModel: getting started, running models, exact mode and precision, the operators, the kernels, how it is checked against ONNX Runtime…" version: "0.3.0-preview.1" --- # onemodel docs Everything about OneModel: getting started, running models, exact mode and precision, the operators, the kernels, how it is checked against ONNX Runtime and the ONNX backend tests, and the reference. OneModel is a preview: 0.3.0-preview.1. ## Start - [Getting started](https://recursiveai.org/onemodel/docs/getting-started/index.md): Install the OneModel preview for .NET 10, run an ONNX model with one call, read its inputs and outputs, and see which kernels it picked. No native… - [Overview](https://recursiveai.org/onemodel/docs/overview/index.md): A neural-network runtime for .NET in pure managed C#, with no native binaries and no dependency beyond the shared framework. It loads ONNX models and our… ## Running models - [Sessions and plans](https://recursiveai.org/onemodel/docs/graph-runtime/index.md): This note says how an imported model becomes something that runs: how operators are described, how a plan for one set of input shapes is made, how memory… - [ONNX import](https://recursiveai.org/onemodel/docs/onnx-import/index.md): How a model file becomes the in-memory graph. A port can follow this text without the C# code. The schema is onnx.proto and the rules are ONNX's IR.md… - [The OneModel format](https://recursiveai.org/onemodel/docs/format/index.md): A model file with the graph and its weights in one place, laid out so the weights can be used straight from a memory map. This text is normative: a reader… - [Graph passes](https://recursiveai.org/onemodel/docs/graph-passes/index.md): Rewrites applied to the compiled graph when a session is created, selected by SessionOptions.OptimizationLevel and the mode. Interface and ownership… ## Exact mode and precision - [Exact mode](https://recursiveai.org/onemodel/docs/exact-mode/index.md): Exact mode gives bit-identical results on every operating system, CPU, instruction set, thread count and runtime (including WebAssembly), so a port to… - [Golden hashes](https://recursiveai.org/onemodel/docs/golden-hashes/index.md): Exact mode (see exact-mode.md) promises the same bits on every OS, CPU, instruction set, thread count and runtime. This file records the check of that… - [Exact transcendentals](https://recursiveai.org/onemodel/docs/exact-transcendentals/index.md): Normative. A port that follows this text reproduces the exact results bit for bit. Source: src/OneModel/Kernels/Exact/Transcendental.cs (the algorithms)… - [Half-precision storage](https://recursiveai.org/onemodel/docs/fp16/index.md): The library computes in binary32. A model stored in binary16 or bfloat16 runs with half-precision storage between operators and binary32 arithmetic inside… - [int8 inference](https://recursiveai.org/onemodel/docs/int8/index.md): Operators, kernels and graph rewrites for 8-bit quantized models. Integer arithmetic is exact, so every int8 kernel gives the same bits on every… ## Operators - [Operators for coverage](https://recursiveai.org/onemodel/docs/ops-coverage/index.md): Semantics the ONNX text leaves open, as settled by ONNX's backend-test data or, where that is silent, by ONNX Runtime 1.31.1's output on crafted inputs.… - [Element-wise operators](https://recursiveai.org/onemodel/docs/ops-elementwise/index.md): Normative notes for the operators in src/OneModel/Graph/Operators/Elementwise/. Semantics come from the ONNX operator specification; where the… - [Normalization and reductions](https://recursiveai.org/onemodel/docs/ops-norms/index.md): Normative for the exact flavour; the fast flavour is the same unless a section says otherwise. Source: src/OneModel/Kernels/Exact/NormKernels.cs… - [Convolution and pooling](https://recursiveai.org/onemodel/docs/ops-conv/index.md): Behaviour of the operators of this family beyond what the ONNX operator text says: which cases are supported, how the open edges were settled, and which… - [Shape and indexing](https://recursiveai.org/onemodel/docs/ops-shape-indexing/index.md): Notes on the operators that move, select or describe elements without computing on them: what each one does where the ONNX text leaves room, and how each… - [RNN, GRU, LSTM](https://recursiveai.org/onemodel/docs/ops-recurrent/index.md): Semantics are those of Operators.md (opsets 7, 14, 22 are one implementation: the attributes and inputs did not change in a way that needs a second one… - [CTC decoding](https://recursiveai.org/onemodel/docs/ctc/index.md): A helper for consumers of recognisers and speech models (not an ONNX operator): RecursiveAI.OneModel.Tensors.Ctc. It turns per-step class scores into… ## Kernels - [Tensors and memory](https://recursiveai.org/onemodel/docs/tensors/index.md): Normative description of the tensor layer: element types, shapes and strides, views, layouts, buffer ownership, version counters, the backward record, the… - [Matrix product kernels](https://recursiveai.org/onemodel/docs/kernels-gemm/index.md): fp32, row-major, C = A * B with A m x k, B k x n and row strides lda, ldb, ldc; convolution in NCHW. Both modes exist for every kernel; the exact mode's… - [Attention and softmax](https://recursiveai.org/onemodel/docs/kernels-attention/index.md): Normative for the exact flavour. Source: src/OneModel/Kernels/Attention.cs, Softmax.cs, Exact/Transcendental*.cs. - [Depthwise convolution](https://recursiveai.org/onemodel/docs/kernels-depthwise/index.md): Applies to a Conv with Groups == Cin (one input channel per group; Cout = Cin * multiplier). Other convolutions are lowered to the matrix product… ## How it is checked - [ONNX backend tests](https://recursiveai.org/onemodel/docs/conformance/index.md): Generated by tools/conformance.sh (it runs OnnxSweep conformance over the frozen ONNX 1.22.0 backend-test data, then tools/conformance_table.py). Nothing… - [Reference models](https://recursiveai.org/onemodel/docs/reference-models/index.md): The seven reference models run through Model.Load(path).Run(...) or Session.Run(...) and are compared with ONNX Runtime's outputs stored by… - [Oracles and reference data](https://recursiveai.org/onemodel/docs/oracles/index.md): ONNX Runtime and PyTorch are used as black boxes: they produce reference outputs, export models and quantize them. Their code is never read, copied or… - [Benchmark method](https://recursiveai.org/onemodel/docs/bench-method/index.md): A row compares two engines on one model file with the same inputs in the same harness: ONNX Runtime (the Microsoft.ML.OnnxRuntime package pinned in… ## Reference - [Docs map](https://recursiveai.org/onemodel/docs/docs-map/index.md): Normative, language-neutral notes live here, one file per topic, so a port can follow them without the C# code. - [Changelog](https://recursiveai.org/onemodel/docs/changelog/index.md): All notable changes are listed here. The format follows Keep a Changelog, and versions follow Semantic Versioning. ## Cookbook - [All recipes](https://recursiveai.org/onemodel/cookbook/index.md): 5 short, task-first recipes with complete code.