FINETUNE ODOR HOW TO
FINETUNE ODOR CODE
FINETUNE ODOR DOWNLOAD

Now that we have the basic class structure, lets define some helper functions for creating the layers. The load_initial_weights function will be used to assign the pretrained weights to our created variables. We could do all in once, but I personally find this a much cleaner way.

In the _init_ function we will parse the input arguments to class variables and call the create function. create () def create ( self ): pass def load_initial_weights ( self ): pass WEIGHTS_PATH = weights_path # Call the create function to build the computational graph of AlexNet WEIGHTS_PATH = 'bvlc_alexnet.npy' else : self. IS_TRAINING = is_training if weights_path = 'DEFAULT' : self. (if bvlc_alexnet.npy is not in the same folder) weights_path: path string, path to the pretrained weights, skip_layer: list of strings, names of the layers you want to reinitialize num_classes: int, number of classes of the new dataset keep_prob: tf.placeholder, for the dropout rate x: tf.placeholder, for the input images Let’s first look onto the model structure as shown in the original paper:Ĭlass AlexNet ( object ): def _init_ ( self, x, keep_prob, num_classes, skip_layer, weights_path = 'DEFAULT' ): ''' But don’t worry, we don’t have to do everything manually. This is the same thing I defined for BatchNormalization in my last blog post but for the entire model. To start finetune AlexNet, we first have to create the so-called “Graph of the Model”.

FINETUNE ODOR DOWNLOAD

Anyway, here you can download the already converted weights. I tried it on my own and it works pretty straight forward.

FINETUNE ODOR CODE

Luckily Caffe to TensorFlow exists, a small conversion tool, to translate any *prototxt model definition from caffe to python code and a TensorFlow model, as well as conversion of the weights. Caffe does, but it’s not to trivial to convert the weights manually in a structure usable by TensorFlow. Unlike VGG or Inception, TensorFlow doesn’t ship with a pretrained AlexNet. I’ll explain most of the steps you need to do, but basic knowledge of TensorFlow and machine/deep learning is required to fully understand everything.

FINETUNE ODOR HOW TO

Although I recommend reading the first part, click here to skip the first part and go directly on how to finetune AlexNet.Īnd next: This is not an introduction neither to TensorFlow nor to finetuning or convolutional networks in general. Basically it is divided into two parts: In the first part I created a class to define the model graph of AlexNet together with a function to load the pretrained weights and in the second part how to actually use this class to finetune AlexNet on a new dataset. Well and I think the main reason for this article is that working on a project like this, helps me to better understand TensorFlow in general.Īfter finishing to write this article I ended up having written another very long post.After some time with Keras, I recently switched to pure TensorFlow and now I want to be able to finetune the same network as previously, but using just TensorFlow. prototxt, the model structure with blobs…) to work with Caffe. But to be honest, I found it quite cumbersome (e.g. In the past, I used mainly Caffe to finetune convolutional networks.As the models get deeper they naturally need more computational time, which in some projects I can’t afford. Another reason is that for a lot of my personal projects AlexNet works quite well and there is no reason to switch to any of the more heavy-weight models to gain probably another.For the AlexNet model, we have to do a bit more on our own. Although the idea behind finetuning is the same, the major difference is, that Tensorflow (as well as Keras) already ship with VGG or Inception classes and include the weights (pretrained on ImageNet). Albeit there exist many How-To’s, most of the newer once are covering finetuning VGG or Inception Models and not AlexNet.Well here are some reasons why I thought it’s worth doing anyway: “AlexNet?” you might say, “So 2012’ish!” you might say. This time about finetuning AlexNet in pure TensorFlow 1.0. The links below in this article are still pointing to the code explained here in this article.Īfter over one year I finally found time and leisure to write my next article. Read my other blogpost for an explanation of this new feature coming with TensorFlows version >= 1.12rc0. I updated the code of the repository to work with TensorFlows new input pipeline.

Finetune odor