Awesome

ssc4onnx

Checker with simple ONNX model structure. Simple Structure Checker for ONNX.

https://github.com/PINTO0309/simple-onnx-processing-tools

Key concept

• Analyzes and displays the structure of huge size models that cannot be displayed by Netron.

1. Setup

1-1. HostPC

### option
$ echo export PATH="~/.local/bin:$PATH" >> ~/.bashrc \
&& source ~/.bashrc

### run
$ pip install -U onnx rich onnxruntime \
&& pip install -U ssc4onnx \
&& python -m pip install onnx_graphsurgeon \
      --index-url https://pypi.ngc.nvidia.com

1-2. Docker

https://github.com/PINTO0309/simple-onnx-processing-tools#docker

2. CLI Usage

$ ssc4onnx -h

usage:
    ssc4onnx [-h]
    -if INPUT_ONNX_FILE_PATH

optional arguments:
  -h, --help
        show this help message and exit.

  -if INPUT_ONNX_FILE_PATH, --input_onnx_file_path INPUT_ONNX_FILE_PATH
        Input onnx file path.

3. In-script Usage

>>> from ssc4onnx import structure_check
>>> help(structure_check)

Help on function structure_check in module ssc4onnx.onnx_structure_check:

structure_check(
    input_onnx_file_path: Union[str, NoneType] = '',
    onnx_graph: Union[onnx.onnx_ml_pb2.ModelProto, NoneType] = None
) -> Tuple[Dict[str, int], int]

    Parameters
    ----------
    input_onnx_file_path: Optional[str]
        Input onnx file path.
        Either input_onnx_file_path or onnx_graph must be specified.
        Default: ''

    onnx_graph: Optional[onnx.ModelProto]
        onnx.ModelProto.
        Either input_onnx_file_path or onnx_graph must be specified.
        onnx_graph If specified, ignore input_onnx_file_path and process onnx_graph.

    Returns
    -------
    op_num: Dict[str, int]
        Num of every op
    model_size: int
        Model byte size

4. CLI Execution

$ ssc4onnx -if deqflow_b_things_opset12_192x320.onnx

5. In-script Execution

from ssc4onnx import structure_check

structure_check(
  input_onnx_file_path="deqflow_b_things_opset12_192x320.onnx",
)

6. Sample

https://github.com/PINTO0309/ssc4onnx/releases/download/1.0.6/deqflow_b_things_opset12_192x320.onnx

https://github.com/PINTO0309/ssc4onnx/assets/33194443/fd6a4aa2-9ed5-492b-82ae-1f8306af5119

7. Reference

1. https://github.com/onnx/onnx/blob/main/docs/Operators.md
2. https://docs.nvidia.com/deeplearning/tensorrt/onnx-graphsurgeon/docs/index.html
3. https://github.com/NVIDIA/TensorRT/tree/main/tools/onnx-graphsurgeon
4. https://github.com/PINTO0309/simple-onnx-processing-tools
5. https://github.com/PINTO0309/PINTO_model_zoo

8. Issues

https://github.com/PINTO0309/simple-onnx-processing-tools/issues