Skip to content
#

m5

Here are 35 public repositories matching this topic...

Benchmarking time-series foundation models (Chronos-Bolt, zero-shot) vs. supervised (PatchTST) and classical (seasonal-naive, Croston) baselines on the M5 Walmart dataset, scored with MASE and WQL. No single model dominates: foundation/deep models win on dense SKUs, classical methods win on the intermittent tail.

  • Updated Jul 19, 2026
  • Python

Add this topic to your repo

To associate your repository with the m5 topic, visit your repo's landing page and select "manage topics."

Learn more