Tensorflow slim pre-trained alexnet

Question

The Tensorflow slim library provides the graph structure for alexnet, however it doesn't seem to provide a pre-trained (on Imagenet) alexnet checkpoint (https://github.com/tensorflow/models/tree/master/research/slim). Is there anywhere a pre-trained alexnet checkpoint for TF-slim is provided?

I'm searching for this as well... did you ever find an alexnet checkpoint @h1234s? — Richard, Apr 28 '17 at 19:18
Me as well. Here's a vanilla TF one: https://github.com/guerzh/tf_weights — txizzle, Jun 29 '17 at 00:16

score 3 · Answer 1 · answered Dec 21 '17 at 11:00

Unlike VGG or Inception, TensorFlow doesn’t ship with a pretrained AlexNet. Caffe does but it's not a trivial task to convert to tensorflow.

Thanks to Frederik Kratzert, he did that job and share here

In summary, you need to:

Download Caffe model and weights
Use the caffe-to-tensorflow to convert the Caffe model to Tensorflow model
Refer the diferences if any between the converted model and the slim model and make the adaptation

For more details, please look here

Tensorflow slim pre-trained alexnet

1 Answers1