Plankifier#
Repo: pegasus-isi/kiso-plankifier-experiment
Components: Chameleon · Chameleon Edge · Pegasus
A deep learning classifier for automated plankton image categorization. Plankton are effective indicators of environmental change in freshwater habitats, but manual microscopic annotation of millions of images is labour-intensive. This experiment runs a Pegasus workflow that applies a deep learning model to classify lake zooplankton images at scale.
The experiment provisions a two-node HTCondor cluster: a cloud node acting as central manager, submit node, and execute node, and an edge node acting as an additional execute node. This edge-to-cloud setup demonstrates how Pegasus can dispatch classification tasks across both cloud and edge resources within a single workflow.
git clone https://github.com/pegasus-isi/kiso-plankifier-experiment
cd kiso-plankifier-experiment
kiso check experiment.yml
kiso up experiment.yml
kiso run experiment.yml
kiso down experiment.yml
See the experiment README for prerequisites, dataset setup, and configuration details.
See also#
Experiment gallery — all experiments