In general, the pre-trained models in OpenVINO are models where training has already occurred, and often have higher, or even cutting-edge accuracy. By using the pre-trained models, one gets to avoid the need for large-scale data collection and long, costly training.
The models further provides knowledge of how to pre-process the inputs and handle the outputs of the network, where you can easily plug these directly into your own application.
In OpenVINO, the Pre-Trained Models includes the Model Zoo, where the Free Model Set contains pre-trained models already converted using the Model Optimizer.
The advantage with the Pr-Trained Models is that one can easily use them directly with the Inference Engine.
Majority of the Pre-Trained Models are either:
- The human detection
- The vehicle detection
- The text detection
- Age detection
- Gender detection
- Emotion recognition
- Or the face detection.
These Pre-Trained Models mostly use the SSD and the MobileNet architectures.
It is however important to note that one can get other Pre-Trained Models from other sources and still use them in the OpenVINO Toolkit too :).
This link shows more case studies on how the OpenVINO Toolkit has been used.
If you have any question or comment, do not hesitate to ask us.
Quote: The moon looks upon many night flowers; the night flowers see but one moon. – Jean Ingelow