OpenVINO™ Training Extensions offers a powerful, low-code solution for transfer learning tailored specifically for computer vision tasks. Whether you’re a seasoned professional or a novice diving into deep learning, this framework simplifies the training, inference, optimizing, and deployment of models. In this blog, we’ll guide you through its key features, installation, and initial usage.

Key Features

• Diverse Task Support: OpenVINO™ Training Extensions can manage a wide range of tasks including:
  • Classification
  • Object Detection
  • Semantic Segmentation
  • Action Recognition
  • Anomaly Recognition
  • Visual Prompting
• Learning Methods: Supports various learning methods including supervised and incremental training.
• Usability Features:
  • Auto-configuration to optimize model settings
  • Support for various dataset formats
  • Distributed training and mixed-precision training for performance
  • Integrated hyper-parameter optimization module (HPO)

Installation

To start using OpenVINO™ Training Extensions, follow the steps provided in the installation guide. Here’s a summary of the installation methods:

Install from PyPI

The easiest way to install is by using pip:

pip install otx[base]

Install from Source

For a local installation, clone the repository and install it in editable mode:

# Use of virtual environment is highly recommended
conda create -n otx_env python=3.10
conda activate otx_env
git clone https://github.com/openvinotoolkit/training_extensions.git
cd training_extensions
pip install -e .[base]  # for zsh: pip install -e .[base]

Quick Start

OpenVINO™ Training Extensions allows for both API-based and CLI-based training. The API provides greater flexibility while the CLI is more straightforward for quick usage.

Training via API

To train using the provided auto-configuration:

from otx.engine import Engine

engine = Engine(data_root='data/path', task='DETECTION')
engine.train()

Training via CLI

For CLI users:

otx train --data_root data/path --task DETECTION

For additional examples, refer to the API Quick-Guide and CLI Guide.

Troubleshooting

While using OpenVINO™ Training Extensions, you may encounter issues or questions. Here are some troubleshooting ideas:

• Dependency Issues: Ensure that all package dependencies are correctly installed.
• Compatibility Errors: Make sure that your Python and library versions meet the specified requirements.
• Dataset Configuration: Double-check dataset paths and formats if errors arise during training.

For more insights, updates, or to collaborate on AI development projects, stay connected with fxis.ai.

Conclusion

At fxis.ai, we believe that such advancements are crucial for the future of AI, as they enable more comprehensive and effective solutions. Our team is continually exploring new methodologies to push the envelope in artificial intelligence, ensuring that our clients benefit from the latest technological innovations.