OpenVINO™ Training Extensions provide a convenient environment to train Deep Learning models and convert them using the OpenVINO™ toolkit for optimized inference.

• Ubuntu 18.04 / 20.04
• Python 3.6+
• OpenVINO™ - for exporting and running models
• CUDA Toolkit 10.2 - for training on GPU

1. Clone repository in the working directory by running the following:

   git clone https://github.com/openvinotoolkit/training_extensions.git
   export OTE_DIR=`pwd`/training_extensions
   

2. Clone Open Model Zoo repository to run demos:

   git clone https://github.com/openvinotoolkit/open_model_zoo --branch develop
   export OMZ_DIR=`pwd`/open_model_zoo
   

3. Install prerequisites by running the following:

   sudo apt-get install python3-pip virtualenv
   

4. Create and activate virtual environment:

   cd training_extensions
   virtualenv venv
   source venv/bin/activate
   

5. Install ote package:

   pip3 install -e ote/
   

After installation, you are ready to train your own models, evaluate and use them for prediction.

• Action Recognition
  • Custom Action Recognition
  • Gesture Recognition
• Instance Segmentation
  • COCO instance segmentation
  • Custom instance segmentation
• Image classification
  • Custom image classification
• Object Detection
  • Custom Object Detection
  • Face Detection
  • Horizontal Text Detection
  • Person Detection
  • Person Vehicle Bike Detection
  • Vehicle Detection
• Text Spotting
  • Alphanumeric Text Spotting

The image classification and object detection models can be compressed by NNCF framework.

See details in the corresponding readme files of the models.

Models that were previously developed can be found here.

Please read the contribution guidelines before starting work on a pull request.

Currently, training, exporting, evaluation scripts for TensorFlow*-based models and the most of PyTorch*-based models from Misc section are exploratory and are not validated.

* Other names and brands may be claimed as the property of others.