densenet264 pretrained network loading problem

hi all i am loading densenet264 as 'net=densenet264' but is gives me the error "Undefined function or variable 'densenet264'."

i have already used it for densenet201 but it work fine

thanks in anticipation

Answer 1

[Prasanna]

Hi Asif,

The issue you're encountering likely stems from the fact that densenet201 is a built-in pretrained model available directly in the Deep Learning Toolbox Model for ResNet Network support package, whereas densenet264 is not included in the supported list of pretrained networks.

If you specifically need DenseNet-264, you will need to either:

• Manually define the architecture using layerGraph and construct the network based on the original DenseNet-264 configuration from the research paper.
• Import a pretrained DenseNet-264 model from an external source like TensorFlow, PyTorch, or ONNX, and convert it using the Deep Learning Toolbox Converter for ONNX Model Format, assuming you can get the model in ONNX format.

In conclusion, densenet264 is not available as a pretrained model , which is why you get the "undefined function" error. You may continue using available networks like densenet201, or construct/import DenseNet-264 manually. For more information, you can refer to:

• Pretrained convolutional neural networks: https://www.mathworks.com/help/deeplearning/ug/pretrained-convolutional-neural-networks.html
• Import network from ONNX: https://www.mathworks.com/help/deeplearning/ref/importnetworkfromonnx.html
• Densenet201: https://www.mathworks.com/help/deeplearning/ref/densenet201.html

Hope this helps!