Skip to main content

Image classification library for easily and quickly deploying models and training classifiers

Project description

Codacy Badge License: MIT GitHub release



TensorImage is an open source library designed to make training and deploying image classification models easy.


  • Cluster training: automatically compare the performance of multiple trainers, speeding up the process of hyperparameter tuning and feature engineering, as there is no need to do it manually

  • Multithreaded training: by default, all training operations are run in 10 threads to make training models faster

  • Built-in image data augmentation operations, which can be used for feature engineering:

    • Image flipping
    • Salt-pepper noise
    • Random brightness
    • Random contrast
    • Random hue
    • Random saturation
    • Gaussian blur
    • Colour filtering
  • Workspace organization: all datasets, trained models, and internal metadata files are stored automatically inside a workspace directory, where you can quickly find any files you need

  • Large-scale image classification: deploy trained models on thousands of images, with predictions for all images being stored in your workspace directory

Upcoming features

  • More data augmentation operations to apply on images:

    • Affine/perspective transformations
    • Random zooming
    • Random cropping
    • Individual pepper and salt noise
    • More image blurring techniques:
      • Median blur
      • Average blur
      • Motion blur
      • Bilateral blur
    • Translation
  • Option to apply different data augmentation operations at once, e.g: instead of only applying gaussian blur, to be able to apply gaussian blur, pepper salt noise and random contrast at once, not uniquely separately

  • Model inference for individual/batches of images for real-time prediction without writing on disk

  • Real-time training from individual/batches of images without reading from disk, automatically training the model from new data, linked to real-time inference without having to store the model in disk (with option to store available)


From the terminal:

$ pip3 install tensorimage


You can view TensorImage's documentation here.


If you are experiencing any errors or bugs, please report them in the issues section or contact us at


If you have any ideas for features that should be added to TensorImage, please feel free to fork TensorImage and open a pull request.


TensorImage is licensed under the MIT license.

Project details

Download files

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

Files for TensorImage, version 2.1.0
Filename, size File type Python version Upload date Hashes
Filename, size TensorImage-2.1.0-py3-none-any.whl (30.4 kB) File type Wheel Python version py3 Upload date Hashes View
Filename, size TensorImage-2.1.0.tar.gz (17.3 kB) File type Source Python version None Upload date Hashes View

Supported by

Pingdom Pingdom Monitoring Google Google Object Storage and Download Analytics Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page