TI deep learning (TIDL) -- 3

1.4. Training As long as layers are supported and parameter constraints are met, existing cafe and TF slim models can be imported. However, these models usually include dense weight matrices. In order to take advantage of some advantages of tidl lib and obtain 3x-4x performance improvement (for convolution layer), it is necessary to use caffe J ...

Added by mikeatrpi on Sun, 03 Oct 2021 02:06:04 +0300

Classic network architecture of deep learning: AlexNet

1, Introduction Alex net was designed by Hinton, the winner of the 2012 ImageNet competition, and his student Alex Krizhevsky. Also after that year, more and deeper neural networks were proposed, such as the excellent VGg and Google lenet. The accuracy of the official data model is 57.1% and the top 1-5 is 80.2%. This is quite excellent for th ...

Added by intermediate on Thu, 30 Sep 2021 01:30:52 +0300