Very simple NCHW and NHWC conversion tool for ONNX. Change to the specified input order for each and every input OP. Also, change the channel order of RGB and BGR. Simple Channel Converter for ONNX.

Overview

scc4onnx

Very simple NCHW and NHWC conversion tool for ONNX. Change to the specified input order for each and every input OP. Also, change the channel order of RGB and BGR. Simple Channel Converter for ONNX.

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

Downloads GitHub PyPI CodeQL

Key concept

  • Allow the user to specify the name of the input OP to change the input order.
  • All number of dimensions can be freely changed, not only 4 dimensions such as NCHW and NHWC.
  • Simply rewrite the input order of the input OP to the specified order and extrapolate Transpose after the input OP so that it does not affect the processing of subsequent OPs.
  • Allows the user to change the channel order of RGB and BGR by specifying options.

1. Setup

1-1. HostPC

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

### run
$ pip install -U onnx \
&& python3 -m pip install -U onnx_graphsurgeon --index-url https://pypi.ngc.nvidia.com \
&& pip install -U scc4onnx

1-2. Docker

### docker pull
$ docker pull pinto0309/scc4onnx:latest

### docker build
$ docker build -t pinto0309/scc4onnx:latest .

### docker run
$ docker run --rm -it -v `pwd`:/workdir pinto0309/scc4onnx:latest
$ cd /workdir

2. CLI Usage

$ scc4onnx -h

usage:
  scc4onnx [-h]
  --input_onnx_file_path INPUT_ONNX_FILE_PATH
  --output_onnx_file_path OUTPUT_ONNX_FILE_PATH
  [--input_op_names_and_order_dims INPUT_OP_NAME ORDER_DIM]
  [--channel_change_inputs INPUT_OP_NAME DIM]
  [--non_verbose]

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

  --input_onnx_file_path INPUT_ONNX_FILE_PATH
      Input onnx file path.

  --output_onnx_file_path OUTPUT_ONNX_FILE_PATH
      Output onnx file path.

  --input_op_names_and_order_dims INPUT_OP_NAME ORDER_DIM
      Specify the name of the input_op to be dimensionally changed and the order of the
      dimensions after the change.
      The name of the input_op to be dimensionally changed can be specified multiple times.

      e.g.
      --input_op_names_and_order_dims aaa [0,3,1,2] \
      --input_op_names_and_order_dims bbb [0,2,3,1] \
      --input_op_names_and_order_dims ccc [0,3,1,2,4,5]

  --channel_change_inputs INPUT_OP_NAME DIM
      Change the channel order of RGB and BGR.
      If the original model is RGB, it is transposed to BGR.
      If the original model is BGR, it is transposed to RGB.
      It can be selectively specified from among the OP names specified
      in --input_op_names_and_order_dims.
      OP names not specified in --input_op_names_and_order_dims are ignored.
      Multiple times can be specified as many times as the number of OP names specified
      in --input_op_names_and_order_dims.
      --channel_change_inputs op_name dimension_number_representing_the_channel
      dimension_number_representing_the_channel must specify the dimension position before
      the change in input_op_names_and_order_dims.
      For example, dimension_number_representing_the_channel is 1 for NCHW and 3 for NHWC.

      e.g.
      --channel_change_inputs aaa 3 \
      --channel_change_inputs bbb 1 \
      --channel_change_inputs ccc 5

  --non_verbose
      Do not show all information logs. Only error logs are displayed.

3. In-script Usage

$ python
>>> from scc4onnx import order_conversion
>>> help(order_conversion)
Help on function order_conversion in module scc4onnx.onnx_input_order_converter:

order_conversion(
  input_op_names_and_order_dims: Union[dict, NoneType] = None,
  channel_change_inputs: Union[dict, NoneType] = None,
  input_onnx_file_path: Union[str, NoneType] = '',
  output_onnx_file_path: Union[str, NoneType] = '',
  onnx_graph: Union[onnx.onnx_ml_pb2.ModelProto, NoneType] = None,
  non_verbose: Union[bool, NoneType] = False
) -> onnx.onnx_ml_pb2.ModelProto

    Parameters
    ----------
    input_onnx_file_path: Optional[str]
        Input onnx file path.
        Either input_onnx_file_path or onnx_graph must be specified.
    
    output_onnx_file_path: Optional[str]
        Output onnx file path.
        If output_onnx_file_path is not specified, no .onnx file is output.
    
    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.
    
    input_op_names_and_order_dims: Optional[dict]
        Specify the name of the input_op to be dimensionally changed and
        the order of the dimensions after the change.
        The name of the input_op to be dimensionally changed
        can be specified multiple times.
    
        e.g.
        input_op_names_and_order_dims = {
            "input_op_name1": [0,3,1,2],
            "input_op_name2": [0,2,3,1],
            "input_op_name3": [0,3,1,2,4,5],
        }
    
    channel_change_inputs: Optional[dict]
        Change the channel order of RGB and BGR.
        If the original model is RGB, it is transposed to BGR.
        If the original model is BGR, it is transposed to RGB.
        It can be selectively specified from among the OP names
        specified in input_op_names_and_order_dims.
        OP names not specified in input_op_names_and_order_dims are ignored.
        Multiple times can be specified as many times as the number
        of OP names specified in input_op_names_and_order_dims.
        channel_change_inputs = {"op_name": dimension_number_representing_the_channel}
        dimension_number_representing_the_channel must specify
        the dimension position after the change in input_op_names_and_order_dims.
        For example, dimension_number_representing_the_channel is 1 for NCHW and 3 for NHWC.
    
        e.g.
        channel_change_inputs = {
            "aaa": 1,
            "bbb": 3,
            "ccc": 2,
        }
    
    non_verbose: Optional[bool]
        Do not show all information logs. Only error logs are displayed.
        Default: False
    
    Returns
    -------
    order_converted_graph: onnx.ModelProto
        Order converted onnx ModelProto

4. CLI Execution

$ scc4onnx \
--input_onnx_file_path crestereo_next_iter2_240x320.onnx \
--output_onnx_file_path crestereo_next_iter2_240x320_ord.onnx \
--input_op_names_and_order_dims left [0,2,3,1] \
--input_op_names_and_order_dims right [0,2,3,1] \
--channel_change_inputs left 1 \
--channel_change_inputs right 1

5. In-script Execution

from scc4onnx import order_conversion

order_converted_graph = order_conversion(
    onnx_graph=graph,
    input_op_names_and_order_dims={"left": [0,2,3,1], "right": [0,2,3,1]},
    channel_change_inputs={"left": 1, "right": 1},
    non_verbose=True,
)

6. Sample

6-1. Transpose only

image

$ scc4onnx \
--input_onnx_file_path crestereo_next_iter2_240x320.onnx \
--output_onnx_file_path crestereo_next_iter2_240x320_ord.onnx \
--input_op_names_and_order_dims left [0,2,3,1] \
--input_op_names_and_order_dims right [0,2,3,1]

image image

6-2. Transpose + RGB<->BGR

image

$ scc4onnx \
--input_onnx_file_path crestereo_next_iter2_240x320.onnx \
--output_onnx_file_path crestereo_next_iter2_240x320_ord.onnx \
--input_op_names_and_order_dims left [0,2,3,1] \
--input_op_names_and_order_dims right [0,2,3,1] \
--channel_change_inputs left 1 \
--channel_change_inputs right 1

image

6-3. RGB<->BGR only

image

$ scc4onnx \
--input_onnx_file_path crestereo_next_iter2_240x320.onnx \
--output_onnx_file_path crestereo_next_iter2_240x320_ord.onnx \
--channel_change_inputs left 1 \
--channel_change_inputs right 1

image

7. Issues

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

You might also like...
Python script for performing depth completion from sparse depth and rgb images using the msg_chn_wacv20. model in ONNX
Python script for performing depth completion from sparse depth and rgb images using the msg_chn_wacv20. model in ONNX

ONNX msg_chn_wacv20 depth completion Python script for performing depth completion from sparse depth and rgb images using the msg_chn_wacv20 model in

A Planar RGB-D SLAM which utilizes Manhattan World structure to provide optimal camera pose trajectory while also providing a sparse reconstruction containing points, lines and planes, and a dense surfel-based reconstruction.
A Planar RGB-D SLAM which utilizes Manhattan World structure to provide optimal camera pose trajectory while also providing a sparse reconstruction containing points, lines and planes, and a dense surfel-based reconstruction.

ManhattanSLAM Authors: Raza Yunus, Yanyan Li and Federico Tombari ManhattanSLAM is a real-time SLAM library for RGB-D cameras that computes the camera

Using Streamlit to host a multi-page tool with model specs and classification metrics, while also accepting user input values for prediction.

Predicitng_viability Using Streamlit to host a multi-page tool with model specs and classification metrics, while also accepting user input values for

Python project to take sound as input and output as RGB + Brightness values suitable for DMX

sound-to-light Python project to take sound as input and output as RGB + Brightness values suitable for DMX Current goals: Get one pixel working: Vary

Feed forward VQGAN-CLIP model, where the goal is to eliminate the need for optimizing the latent space of VQGAN for each input prompt
Feed forward VQGAN-CLIP model, where the goal is to eliminate the need for optimizing the latent space of VQGAN for each input prompt

Feed forward VQGAN-CLIP model, where the goal is to eliminate the need for optimizing the latent space of VQGAN for each input prompt. This is done by

An executor that loads ONNX models and embeds documents using the ONNX runtime.

ONNXEncoder An executor that loads ONNX models and embeds documents using the ONNX runtime. Usage via Docker image (recommended) from jina import Flow

ONNX Runtime Web demo is an interactive demo portal showing real use cases running ONNX Runtime Web in VueJS.

ONNX Runtime Web demo is an interactive demo portal showing real use cases running ONNX Runtime Web in VueJS. It currently supports four examples for you to quickly experience the power of ONNX Runtime Web.

A repository that shares tuning results of trained models generated by TensorFlow / Keras. Post-training quantization (Weight Quantization, Integer Quantization, Full Integer Quantization, Float16 Quantization), Quantization-aware training. TensorFlow Lite. OpenVINO. CoreML. TensorFlow.js. TF-TRT. MediaPipe. ONNX. [.tflite,.h5,.pb,saved_model,tfjs,tftrt,mlmodel,.xml/.bin, .onnx] ONNX-GLPDepth - Python scripts for performing monocular depth estimation using the GLPDepth model in ONNX
ONNX-GLPDepth - Python scripts for performing monocular depth estimation using the GLPDepth model in ONNX

ONNX-GLPDepth - Python scripts for performing monocular depth estimation using the GLPDepth model in ONNX

Releases(1.0.5)
  • 1.0.5(Sep 9, 2022)

    • Add short form parameter
      $ scc4onnx -h
      
      usage:
        scc4onnx [-h]
        -if INPUT_ONNX_FILE_PATH
        -of OUTPUT_ONNX_FILE_PATH
        [-ioo INPUT_OP_NAME ORDER_DIM]
        [-cci INPUT_OP_NAME DIM]
        [-n]
      
      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.
      
        -of OUTPUT_ONNX_FILE_PATH, --output_onnx_file_path OUTPUT_ONNX_FILE_PATH
            Output onnx file path.
      
        -ioo INPUT_OP_NAMES_AND_ORDER_DIMS INPUT_OP_NAMES_AND_ORDER_DIMS, --input_op_names_and_order_dims INPUT_OP_NAMES_AND_ORDER_DIMS INPUT_OP_NAMES_AND_ORDER_DIMS
            Specify the name of the input_op to be dimensionally changed and the order of the
            dimensions after the change.
            The name of the input_op to be dimensionally changed can be specified multiple times.
      
            e.g.
            --input_op_names_and_order_dims aaa [0,3,1,2] \
            --input_op_names_and_order_dims bbb [0,2,3,1] \
            --input_op_names_and_order_dims ccc [0,3,1,2,4,5]
      
        -cci CHANNEL_CHANGE_INPUTS CHANNEL_CHANGE_INPUTS, --channel_change_inputs CHANNEL_CHANGE_INPUTS CHANNEL_CHANGE_INPUTS
            Change the channel order of RGB and BGR.
            If the original model is RGB, it is transposed to BGR.
            If the original model is BGR, it is transposed to RGB.
            It can be selectively specified from among the OP names specified
            in --input_op_names_and_order_dims.
            OP names not specified in --input_op_names_and_order_dims are ignored.
            Multiple times can be specified as many times as the number of OP names specified
            in --input_op_names_and_order_dims.
            --channel_change_inputs op_name dimension_number_representing_the_channel
            dimension_number_representing_the_channel must specify the dimension position before
            the change in input_op_names_and_order_dims.
            For example, dimension_number_representing_the_channel is 1 for NCHW and 3 for NHWC.
      
            e.g.
            --channel_change_inputs aaa 3 \
            --channel_change_inputs bbb 1 \
            --channel_change_inputs ccc 5
      
        -n, --non_verbose
            Do not show all information logs. Only error logs are displayed.
      

    Full Changelog: https://github.com/PINTO0309/scc4onnx/compare/1.0.4...1.0.5

    Source code(tar.gz)
    Source code(zip)
  • 1.0.4(May 25, 2022)

  • 1.0.3(May 15, 2022)

  • 1.0.2(May 10, 2022)

  • 1.0.1(Apr 19, 2022)

  • 1.0.0(Apr 18, 2022)

Owner
Katsuya Hyodo
Hobby programmer. Intel Software Innovator Program member.
Katsuya Hyodo
OpenLT: An open-source project for long-tail classification

OpenLT: An open-source project for long-tail classification Supported Methods for Long-tailed Recognition: Cross-Entropy Loss Focal Loss (ICCV'17) Cla

Ming Li 37 Sep 15, 2022
Contrastive Language-Image Pretraining

CLIP [Blog] [Paper] [Model Card] [Colab] CLIP (Contrastive Language-Image Pre-Training) is a neural network trained on a variety of (image, text) pair

OpenAI 11.5k Jan 08, 2023
Official implementation of the paper 'High-Resolution Photorealistic Image Translation in Real-Time: A Laplacian Pyramid Translation Network' in CVPR 2021

LPTN Paper | Supplementary Material | Poster High-Resolution Photorealistic Image Translation in Real-Time: A Laplacian Pyramid Translation Network Ji

372 Dec 26, 2022
Crawl & visualize ICLR papers and reviews

Crawl and Visualize ICLR 2022 OpenReview Data Descriptions This Jupyter Notebook contains the data crawled from ICLR 2022 OpenReview webpages and thei

Federico Berto 75 Dec 05, 2022
The implementation our EMNLP 2021 paper "Enhanced Language Representation with Label Knowledge for Span Extraction".

LEAR The implementation our EMNLP 2021 paper "Enhanced Language Representation with Label Knowledge for Span Extraction". See below for an overview of

杨攀 93 Jan 07, 2023
Image Captioning on google cloud platform based on iot

Image-Captioning-on-google-cloud-platform-based-on-iot - Image Captioning on google cloud platform based on iot

Shweta_kumawat 1 Jan 20, 2022
A deep learning framework for historical document image analysis

DIVA-DAF Description A deep learning framework for historical document image analysis. How to run Install dependencies # clone project git clone https

9 Aug 04, 2022
C3DPO - Canonical 3D Pose Networks for Non-rigid Structure From Motion.

C3DPO: Canonical 3D Pose Networks for Non-Rigid Structure From Motion By: David Novotny, Nikhila Ravi, Benjamin Graham, Natalia Neverova, Andrea Vedal

Meta Research 309 Dec 16, 2022
Package to compute Mauve, a similarity score between neural text and human text. Install with `pip install mauve-text`.

MAUVE MAUVE is a library built on PyTorch and HuggingFace Transformers to measure the gap between neural text and human text with the eponymous MAUVE

Krishna Pillutla 182 Jan 02, 2023
The code release of paper 'Domain Generalization for Medical Imaging Classification with Linear-Dependency Regularization' NIPS 2020.

Domain Generalization for Medical Imaging Classification with Linear Dependency Regularization The code release of paper 'Domain Generalization for Me

Yufei Wang 56 Dec 28, 2022
MazeRL is an application oriented Deep Reinforcement Learning (RL) framework

MazeRL is an application oriented Deep Reinforcement Learning (RL) framework, addressing real-world decision problems. Our vision is to cover the complete development life cycle of RL applications ra

EnliteAI GmbH 222 Dec 24, 2022
FairEdit: Preserving Fairness in Graph Neural Networks through Greedy Graph Editing

FairEdit Relevent Publication FairEdit: Preserving Fairness in Graph Neural Networks through Greedy Graph Editing

5 Feb 04, 2022
Official PyTorch implementation of the paper "TEMOS: Generating diverse human motions from textual descriptions"

TEMOS: TExt to MOtionS Generating diverse human motions from textual descriptions Description Official PyTorch implementation of the paper "TEMOS: Gen

Mathis Petrovich 187 Dec 27, 2022
A Number Recognition algorithm

Paddle-VisualAttention Results_Compared SVHN Dataset Methods Steps GPU Batch Size Learning Rate Patience Decay Step Decay Rate Training Speed (FPS) Ac

1 Nov 12, 2021
Implementation of Bagging and AdaBoost Algorithm

Bagging-and-AdaBoost Implementation of Bagging and AdaBoost Algorithm Dataset Red Wine Quality Data Sets For simplicity, we will have 2 classes of win

Zechen Ma 1 Nov 01, 2021
CUAD

Contract Understanding Atticus Dataset This repository contains code for the Contract Understanding Atticus Dataset (CUAD), a dataset for legal contra

The Atticus Project 273 Dec 17, 2022
Contrastive Learning Inverts the Data Generating Process

Official code to reproduce the results and data presented in the paper Contrastive Learning Inverts the Data Generating Process.

71 Nov 25, 2022
Training deep models using anime, illustration images.

animeface deep models for anime images. Datasets anime-face-dataset Anime faces collected from Getchu.com. Based on Mckinsey666's dataset. 63.6K image

Tomoya Sawada 61 Dec 25, 2022
Very large and sparse networks appear often in the wild and present unique algorithmic opportunities and challenges for the practitioner

Sparse network learning with snlpy Very large and sparse networks appear often in the wild and present unique algorithmic opportunities and challenges

Andrew Stolman 1 Apr 30, 2021
Efficient and Accurate Arbitrary-Shaped Text Detection with Pixel Aggregation Network

Efficient and Accurate Arbitrary-Shaped Text Detection with Pixel Aggregation Network Paddle-PANet 目录 结果对比 论文介绍 快速安装 结果对比 CTW1500 Method Backbone Fine

7 Aug 08, 2022