> For the complete documentation index, see [llms.txt](https://docs.binaexperts.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.binaexperts.com/deployment/deployment/legacy.md).

# Legacy

Legacy deployment refers to the traditional and older method of deploying artificial intelligence models for use in production environments. This approach is typically used when pre-trained models need to be deployed and utilized in production systems. Below are the details regarding development and deployment using the legacy approach:

1. **Development**:
   * During the development phase, the AI model is trained and optimized using training data.
   * This phase involves selecting and training the appropriate model architecture, tuning parameters, applying training techniques, and evaluating model performance.
2. **Preparation for Deployment**:
   * After training the model, it needs to be converted to an executable format and prepared for deployment.
   * This involves converting the model to standard formats such as TensorFlow SavedModel or ONNX and requires advanced configuration for execution in different environments.
3. **Deployment**:
   * In this stage, the prepared model is deployed on servers or local systems.
   * It includes installing and configuring the necessary systems for execution, transferring the model to the production environment, and running the model to respond to input requests.
4. **Maintenance and Monitoring**:
   * After deployment, the model needs to be maintained and its performance monitored.
   * This includes monitoring the model's performance, tracking resource consumption, troubleshooting, and necessary updates for improving performance.

In summary, in legacy deployment, pre-trained models are executed on servers or local systems to function as part of production systems and provide services.