Skip to main content

This package help you visualize the ONNX model graph. Client-Server based architecture lets you share the model, using just a url instead of sharing the entire model.

Project description

ONNX Visualizer

This package help you visualize the ONNX model graph. Client-Server based architecture lets you share the model, using just a url instead of sharing the entire model.

Why not netron?

Netron is a viewer for neural network, deep learning and machine learning models. Netron also supports more formats than just ONNX.

But the problem with netron, you can't visualize the models in remote / virtual machines environments, where most of the time GUI is not given. So the only way is you need to download and run the netron locally or use the web app, where model parsing happens in the web browser, which is super slow.

This package parses the onnx model and serves only required data, using the server-client model, which is highly bandwidth effecient.

Usage

Install

pip install onnx-vis

Visualizing

Run the following command to visualize the model

python3 -m onnx_vis <path to the onnx file> -p <port to run the server>

Example:

 python3 -m onnx_vis resnet18.onnx -p 63325

you can access the visualization on http://127.0.0.1:63325

TODO

  • Make Visualization graph of ONNX Nodes
  • Add node attributes like kernels shape, bias shape to node metadata
  • Add Sidebar to get more detailed information about nodes
  • Add search and span to node feature

Contribution & Issues

Contirbution of any kind PR's, Discussions etc. is very much appreciated :). If you face any issues feel free to open a issue.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

onnx_vis-1.0.7.tar.gz (3.2 kB view details)

Uploaded Source

File details

Details for the file onnx_vis-1.0.7.tar.gz.

File metadata

  • Download URL: onnx_vis-1.0.7.tar.gz
  • Upload date:
  • Size: 3.2 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.2 CPython/3.10.6

File hashes

Hashes for onnx_vis-1.0.7.tar.gz
Algorithm Hash digest
SHA256 261e0a0e89ebeb57b8ad03bda760ba028f6aa64bca5196000cac43fb04f271bb
MD5 a204901e5ff111a4f1b01e42f60e59a4
BLAKE2b-256 32323a0545f8d9f2e1c236d891978dd2238d2ecb52d34635e47ae6de2d615eeb

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page