[MLOPS] From experiment management to model serving and back. A complete usecase, step-by-step!
The recording of our talk at the MLOps World summit. This talk covers a complete example, starting from experiment management and data versioning, building up into a pipeline and finally deploying using ClearML serving with drift monitoring. We then induce artifical drift to trigger the monitoring alerts and go back down the chain to quickly retrain a model and deploy it using canary deployment.
💻 Get a server:
Get started by using our free tier servers: https://app.clear.ml
OR by hosting your own: https://github.com/allegroai/clearml-server
📄 Documentation on Fundamentals:
Get Started: https://clear.ml/docs/latest/docs/getting_started/ds/ds_first_steps
Pip Package / ClearML SDK: https://clear.ml/docs/latest/docs/clearml_sdk
ClearML Data: https://clear.ml/docs/latest/docs/clearml_data/clearml_data
ClearML Agent: https://clear.ml/docs/latest/docs/clearml_agent
Supported libraries for automagical integration: https://clear.ml/docs/latest/docs/integrations/libraries
✨ Follow us and star us!
Github page: https://github.com/allegroai/clearml
Twitter: https://twitter.com/clearmlapp
Slack Channel: https://join.slack.com/t/allegroai-trains/shared_invite/zt-c0t13pty-aVUZZW1TSSSg2vyIGVPBhg
LinkedIn: https://www.linkedin.com/company/clearml