gaohuang/SnapshotEnsemble

Snapshot Ensembles in Torch (Snapshot Ensembles: Train 1, Get M for Free)

46
/ 100
Emerging

When training neural networks for tasks like image classification, you typically train one model. This technique allows you to get multiple distinct, high-performing models (an ensemble) from a single training run, which can then be combined to improve prediction accuracy. It takes your standard neural network training setup and outputs an ensemble of models ready for use. This is useful for machine learning engineers and researchers focused on building robust predictive systems.

189 stars. No commits in the last 6 months.

Use this if you want to improve the accuracy and robustness of your neural network predictions without the cost of training multiple models independently.

Not ideal if you are a beginner looking for a simple, out-of-the-box solution without diving into custom training loop modifications.

deep-learning model-ensembling neural-networks model-optimization
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 20 / 25

How are scores calculated?

Stars

189

Forks

33

Language

Lua

License

BSD-3-Clause

Last pushed

May 16, 2017

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/gaohuang/SnapshotEnsemble"

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