Pretrained Deep Learning Models
Take advantage of model architectures developed by the deep learning research community. Popular models offer a robust architecture and skip the need to start from scratch.
Take advantage of model architectures developed by the deep learning research community. Popular models offer a robust architecture and skip the need to start from scratch.
There are many pretrained models to choose from, and each model will have tradeoffs:
Explore tradeoffs between models in the following sections.
To import any model into MATLAB, the structure is:
>> net = networkname
i.e.
>> net = alexnet >> net = resnet50
If the model is not already downloaded, a link will be provided to download the model in MATLAB.
You can iterate on these models quickly and try out different settings such as data preprocessing steps and training options. Once you have a feel for which settings work well, try a more accurate network to see if it improves your results.
Explore examples:
Explore models that are highly effective for image-based workflows, such as image classification, object detection, and semantic segmentation.
For Semantic Segmentation workflows:
Using any of the predefined network architectures provides a convenient starting point for training semantic segmentation networks. These are commonly used layer architectures for semantic segmentation problems:
See more on creating a semantic segmentation network using Deeplab v3+ here.
Deploy to low-compute, low-power devices such as Raspberry Pi or FPGAs, which requires models with a low memory footprint.
See related topics:
Explore examples:
Looking for models from other frameworks? Use ONNX, TensorFlow-Keras, and Caffe importers to import any network into MATLAB.
Explore examples:
See how to import pretrained Keras layers, and replace unsupported layers with custom layers.