Hi everyone!
I’m working on KaggleFleet, an open-source tool designed to make running Kaggle notebooks more convenient for ML researchers and research teams.
When working on experiments with multiple co-authors, managing notebook execution across different Kaggle accounts can involve a lot of repetitive manual work.
Switching accounts, pushing notebooks, checking execution status, and collecting outputs can interrupt the research workflow.
Kaggle gives every account a free weekly GPU quota. A research group with several
co-authors has several quotas, but using them means logging in and out of Kaggle,
copying notebooks around, and tracking who ran what. KaggleFleet does that part for you:
--json output and meaningful exit codes, and AGENTS.md`` kaggle CLI. The tool itself is pure Python standard library.
pipx install git+https://github.com/towhaEL/KaggleFleet.git
Or
pip install kaggle-fleet
kfleet doctor
kfleet account login --label “My Account”
kfleet notebook list
kfleet run push train.ipynb --watch The project is open source under the MIT license.
I’m particularly interested in feedback from people working on deep learning experiments, model training, and collaborative ML research.
What features would you expect from a tool like this?