Eliot Andres blog
http://blog.ndres.me/
Recent content on Eliot Andres blogHugo -- gohugo.ioen-usWhile the models are training: a deep learning blogWed, 07 Jun 2017 23:02:15 +0200Converting a Caffe model to Tensorflow
http://blog.ndres.me/post/convert-caffe-to-tensorflow/
Wed, 07 Jun 2017 23:02:15 +0200http://blog.ndres.me/post/convert-caffe-to-tensorflow/Converting a Caffe model to Tensorflow The Caffe Model Zoo is an extraordinary place where reasearcher share their models. Caffe is an awesome framework, but you might want to use Tensorflow instead. In this blog post, I’ll show you how to convert the Places 365 model to Tensorflow.
Using Caffe-Tensorflow to convert your model Your best bet is to use the awesome caffe-tensorflow. This project takes a prototxt file as an input and converts it to a python file so you can use the model with Tensorflow.Faster inference in Tensorflow using XLA.
http://blog.ndres.me/post/faster-inference-tensorflow/
Mon, 08 May 2017 14:52:12 +0200http://blog.ndres.me/post/faster-inference-tensorflow/About inference Using neural networks is primarily made of 2 phases: training your model and using it. The later part can also be called inference, forward pass or evaluation.
For most researchers, most of the time is used by training : they have to retrain using different architectures or different parameters. However, if you are using deep learning in production, inference is the most critical step.
What is XLA During the tensorflow summit 2017, the Tensorflow team introduced XLA (Accelerated Linear Algebra).Turn any Jupyter notebook into a RESTful Api
http://blog.ndres.me/post/jupyter-notebook-rest-api/
Sun, 07 May 2017 16:25:35 +0200http://blog.ndres.me/post/jupyter-notebook-rest-api/Jupyter Notebooks are pretty awesome. They allow you to prototype and experiment with ease. But did you know that you can also use them as a backend ?
In this tutorial, I’ll show you how to make a simple api that converts radians to degrees. We should end up with something like this:
$ curl "http://serverIp:8888/convert?angle=180" #{"convertedAngle": 3.141592653589793} Installation Assuming that jupyter notebook is already installed, first install the kernel_gateway: