A mirror of public ML datasets and open model weights. This node is in Riga.

What is here

DirectoryContents
datasets/MNIST, WikiText-2, Tiny Shakespeare
models/all-MiniLM-L6-v2 (sentence embeddings)

We copy only material whose license allows redistribution and keep that license next to the files. Every directory has a SHA256SUMS; run sha256sum -c SHA256SUMS after a download.

Nodes

HostCity
nl.ds-ml.techAmsterdam
se.ds-ml.techStockholm
lv.ds-ml.techRiga
fi.ds-ml.techHelsinki

The four nodes sync every night and serve the same tree. Use the one closest to you.

Downloading

wget -c https://lv.ds-ml.tech/pub/README.txt

Big shards come in 2 GB parts. If a download breaks, run the same wget -c again and it continues where it stopped.