Rusheel86/preflight

Pre-flight checks for PyTorch pipelines. Catch silent failures before they waste your GPU.

28
/ 100
Experimental

Before you train a PyTorch deep learning model, this tool helps you quickly check for common data issues that lead to 'garbage in, garbage out.' It takes your data setup (dataloader) and optionally your model and loss function, then reports on potential problems like invalid data values (NaNs), incorrect image channel order, or imbalanced classes. This is for machine learning engineers and researchers who develop and train deep learning models using PyTorch.

Use this if you want to catch silent data and model setup issues in your PyTorch pipeline that would otherwise waste expensive GPU time and produce a useless model.

Not ideal if you need a comprehensive, continuous monitoring and validation platform for your entire ML lifecycle, or if you're not working with PyTorch deep learning models.

deep-learning pytorch model-training data-validation ml-workflow
No Package No Dependents
Maintenance 13 / 25
Adoption 6 / 25
Maturity 9 / 25
Community 0 / 25

How are scores calculated?

Stars

18

Forks

Language

Python

License

MIT

Last pushed

Mar 15, 2026

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/mlops/Rusheel86/preflight"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.