layout: true --- # Sequence Learning ## Toolkits Korbinian Riedhammer --- # [Theano](http://www.deeplearning.net/software/theano/) ![screenshot](/sequence-learning/09-toolkits/tk-theano.png) --- # [TensorFlow](https://www.tensorflow.org) ![screenshot](/sequence-learning/09-toolkits/tk-tf.png) --- # [Microsoft Cognitive Toolkit](https://docs.microsoft.com/en-us/cognitive-toolkit/) ![screenshot](/sequence-learning/09-toolkits/tk-cntk.png) --- # [Torch](http://torch.ch/) ![screenshot](/sequence-learning/09-toolkits/tk-torch.png) --- # [MXNet](http://mxnet.incubator.apache.org/) ![screenshot](/sequence-learning/09-toolkits/tk-mxnet.png) --- # [Caffe](http://caffe.berkeleyvision.org/)_ ![screenshot](/sequence-learning/09-toolkits/tk-caffe.png) --- # [Keras](https://keras.io) ![screenshot](/sequence-learning/09-toolkits/tk-keras.png) --- # [deeplearning4j](https://deeplearning4j.org/) ![screenshot](/sequence-learning/09-toolkits/tk-dl4j.png) --- # [Core ML](https://developer.apple.com/machine-learning/) ![screenshot](/sequence-learning/09-toolkits/tk-coreml.png) --- # Exercise Pick two. - What are pros and cons? - Which one would you choose, and why? - Work out a basic example that gets you started: [admission.asc](09-toolkits/admission.asc) is a tiny classification task, the last column is the label