onnx-ir 0.21.0

ONNX-IR is a pure Rust library for parsing ONNX models into an intermediate representation that can be used to generate code for various ML/DL frameworks
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
108
109
110
111
112
113
114

#!/usr/bin/env -S uv run
# /// script
# dependencies = [
#   "onnx>=1.15.0",
#   "numpy>=1.24.0",
# ]
# ///

"""
Generate ONNX model with branching (multiple consumers).

Tests:
- Single node output consumed by multiple nodes
- Reference counting works correctly
- Rewiring preserves all connections
- Constants aren't incorrectly removed when shared
"""

import onnx
from onnx import helper, TensorProto
import numpy as np


def create_branching_model():
    """Create model where one node's output feeds multiple consumers."""

    # Input
    input_tensor = helper.make_tensor_value_info("input", TensorProto.FLOAT, [1, 4])

    # Outputs (3 different outputs from branching)
    output1 = helper.make_tensor_value_info("output1", TensorProto.FLOAT, [1, 4])
    output2 = helper.make_tensor_value_info("output2", TensorProto.FLOAT, [1, 4])
    output3 = helper.make_tensor_value_info("output3", TensorProto.FLOAT, [1, 4])

    # Initializers
    const1 = helper.make_tensor(
        name="const1",
        data_type=TensorProto.FLOAT,
        dims=[1, 4],
        vals=np.array([[1.0, 2.0, 3.0, 4.0]], dtype=np.float32).flatten().tobytes(),
        raw=True,
    )

    const2 = helper.make_tensor(
        name="const2",
        data_type=TensorProto.FLOAT,
        dims=[1, 4],
        vals=np.array([[0.5, 0.5, 0.5, 0.5]], dtype=np.float32).flatten().tobytes(),
        raw=True,
    )

    # Create nodes with branching structure
    #
    # input → relu → [branch to 3 consumers]
    #                  ├→ add(const1) → output1
    #                  ├→ mul(const2) → output2
    #                  └→ abs → output3
    #
    nodes = [
        # Common node whose output is consumed by multiple nodes
        helper.make_node("Relu", ["input"], ["relu_out"], name="relu"),
        # Consumer 1: Add
        helper.make_node("Add", ["relu_out", "const1"], ["output1"], name="add"),
        # Consumer 2: Multiply
        helper.make_node("Mul", ["relu_out", "const2"], ["output2"], name="mul"),
        # Consumer 3: Abs
        helper.make_node("Abs", ["relu_out"], ["output3"], name="abs"),
    ]

    # Create the graph
    graph = helper.make_graph(
        nodes,
        "branching_model",
        [input_tensor],
        [output1, output2, output3],
        initializer=[const1, const2],
    )

    # Create the model
    model = helper.make_model(
        graph, producer_name="onnx-ir-test", opset_imports=[helper.make_opsetid("", 16)]
    )

    # Check the model
    onnx.checker.check_model(model)

    return model


def main():
    """Generate and save the ONNX model."""
    model = create_branching_model()

    # Save the model
    output_path = "../fixtures/branching.onnx"
    onnx.save(model, output_path)
    print(f"Model saved to {output_path}")

    # Print model info
    print(f"\nModel info:")
    print(f"  Opset version: {model.opset_import[0].version}")
    print(f"  Inputs: {[inp.name for inp in model.graph.input]}")
    print(f"  Outputs: {[out.name for out in model.graph.output]}")
    print(f"  Nodes: {len(model.graph.node)}")
    for node in model.graph.node:
        print(
            f"    - {node.op_type} ({node.name}): {list(node.input)} → {list(node.output)}"
        )
    print(f"\n  Branching structure:")
    print(f"    relu_out is consumed by: add, mul, abs")


if __name__ == "__main__":
    main()