We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Python-based ML and HPC workflows in the Cloud for science and engineering I PyData Chicago 2022
Python-based machine learning and HPC workflows in the cloud for science and engineering, accelerating simulations and improving visualization with Parcel, a flexible workflow tool for complex applications.
- Python-based workflows in the cloud enable scientists to accelerate complex simulations and improve visualization of results
- Parallel Works’ Parcel package simplifies workflow management and parallelization for complex applications
- Large-scale simulations require concurrent execution of multiple models, optimizing hyperparameters, and managing resources
- Cloud providers offer virtual clusters with linked nodes, optimizing performance and abstraction for users
- Parcel can run arbitrary scripts, executables, and Docker containers, making it a flexible workflow tool
- The platform provides intuitive graphical user interfaces for configuring resources, orchestrating apps, and tracking workflow progress
- Model-agnostic workflow framework enables flexible integration of various models and apps, streamlining engineering design optimization and scientific simulations
- Multi-perspective visualization can be achieved by chaining multiple apps together, allowing for real-time insights
- Python’s flexibility and popularity make it an ideal language for developing and managing workflows
- Cloud-based workflows enable researchers to access massive computing resources, accelerate discovery, and publish results faster