This repository contains public examples for the SKIL Platform. It contains basic tutorials and guides, as well as complete SKIL applications and notebooks that can be imported and run from SKIL.
For each application shown here you need a running SKIL instance. SKIL CE is free and can be downloaded and installed in a few simple steps. You'll find more about the specific requirements of each application in the respective folders.
Applications and tutorials
- CLI tutorial: A four-step, lightning introduction to the SKIL command line interface. You'll analyze the input data, deploy a transform process, then deploy a model and finish by testing SKIL's REST client. (Command line & Java)
- [Clinical LSTM application]((https://github.com/SkymindIO/SKIL_Examples/tree/master/Clinical-LSTM-app): This example shows how to train a recurrent neural network written with DL4J on electronic health record data and deploy it with SKIL. (Java)
- SKIL deployment with Docker: This example shows how import and deploy a TensorFlow model to SKIL, all within a Docker container. (Docker & Python)
- Fraud detection application: In this application you'll learn how to build an anomaly detection system to recognize fraudulent behaviour. (Python notebook)
- CIFAR model deployment: This basic workflow example shows you how to deploy a Keras model trained on the CIFAR dataset. (Python).
- Salesforce app: Salesforce SKIL application
- Sequence classification app: This application shows you a detailed example of a sequence classification problem on synthetic control chart time series. (Java)
- Object detection app: In this application you'll learn how the You only look once (YOLO) model can be used for real-time object detection within SKIL. (Java)
All notebooks found in the
notebook folder of this repository contain Zeppelin notebooks in JSON format that can be imported into any SKIL experiment.
- Training & deploying Keras and DL4J models in the same notebook: In this end-to-end example you'll see how to mix and match several deep learning frameworks (and programming languages) in the same Zeppelin notebook. (Python & Scala)
- Importing a Keras model and training it on Spark: This example will show you how to take a Keras model of your choice, import it into DL4J in one line of code, wrap it into a Spark model and do distributed training with it. (Python & Scala)
- Keras model deployment: Deploy a simple Keras model from the SKIL UI. (Python)
- TensorFlow model deployment: Deploy a TensorFlow model with SKIL. (Python)
- Deploying a KNN model: An example showing you how to deploy a K-nearest-neighbor model with SKIL. (Scala)
- [Running a Spark training job with SKIL]((https://github.com/SkymindIO/SKIL_Examples/blob/master/notebooks/spark_example.json): This notebook shows you how to scale-out your DL4J deep learning model on Apache Spark. (Scala)
- UCI sequence classification example: The notebook corresponding to the above listed sequence classification example. (Scala)