Democratizing ML Pipelines using StreamPipes
summary
An open source IoT toolbox that wires sensors, cameras and brokers into a drag‑and‑drop pipeline editor, running as Docker microservices with 30+ adapters (MQTT, OPC UA, Kafka, ROS, REST). It demonstrates no‑code ML: upload a TensorFlow/H2O model or run automated ML (forecasting, sliding‑window classification and session classification) on labeled data from the data explorer, train jobs take about two hours on the demo set, and the trained model deploys in pipeline to classify sensor states and detect objects in live images.
AI-written summary · about
chapters (11)
- 0:00Introduction by hosts
- 0:50Presenters introduction
- 1:10What is StreamPipes
- 3:59Industrial IoT use cases
- 5:02How StreamPipes works
- 7:11StreamPipes components
- 8:16Pipeline editor overview
- 12:49Demo: Using ML models
- 23:43Sneak peek: Training models
- 28:20Demo: Building ML models
- 34:07Q&A session
related talks
36mDemocratization of ML Pipelines and Bringing ML Workflows to Heterogeneous...
33mDeploying and Scaling Spark ML and Tensorflow AI Models
57mMachine Learning Data Pipelines
52mMachine Learning Data Pipelines
45mWorkshop: Deploy and Monitor ML Pipelines with Open Source and Free Applications
43mSimplifying Real-Time ML Pipelines with Quix Streams: An Open Source Python Library for ML Engineers
29mTeam Teardown: MLOps in the Cloud for IoT
13mMLOps Pipelines to Deliver Tf.Lite Models to Distributed Compute Pools at the Edge