SAVRN Model Hub
AI Training Datasets
Each dataset with its card, its structure (every split, row count and column type), its files, its license, and the models that disclose training on it.
11,122Models
3,164Datasets
285Papers
6,792Publishers
11,852Sourced relationships
Updated 2026-10-18 · How the library is built
3,164 datasets, sorted by most downloaded.
Research snapshot of official legislation collected from e-qanun.az downloadDetailPdf. Not legal advice. Official gazettes / government portals prevail over this corpus. - data/laws.parquet — one row per instrument (id, title, full text, dates, status, identifiers, article count, source URL). - data/articles.parquet — one row per article/section (lawid, article number, title, text, source URL). - scrapers/collectaz.py — research collector used to build this snapshot. - scrapers/common.py — shared collector helpers. - scrapers/worldlib.py — shared fetch (live official URL, Wayback fallback). - scrapers/archivefallbacks.py — Wayback / Common Crawl CDX helpers. Official government / gazette…
Dataset · Text generation
midi-tokens-dyadic-tu0.01-vb32-mts1.0-vocab512-performances-aug-more
Pre-tokenized MIDI pieces for IsoFLOP scaling-law runs. Each row is one full piece (no time-windowing); training crops sequences from packed token bins. The source column is the original piece metadata as JSON so a row can be traced back to its EPR Labs source dataset. Based on MIDI datasets gathered by EPR Labs. - maxtimestep: 1.0 - nvelocitybins: 32 - timeunit: 0.01 The repository also stores midicodec.json and Hugging Face tokenizer files so token ids can be rebuilt exactly. Train pieces only (more setting). Validation and test are unaugmented Maestro. - maestro-sustain-v2 - giant-midi-sustain-v2 - pijamia-midi-v1 - atepp-1.1-sustain-v2 - music-net - piast-midi - aria-pruned-base-v2…
Research snapshot of official legislation collected from documents.gov.lk / parliament.lk. Not legal advice. Official gazettes / government portals prevail over this corpus. - data/laws.parquet — one row per instrument (id, title, full text, dates, status, identifiers, article count, source URL). - data/articles.parquet — one row per article/section (lawid, article number, title, text, source URL). - scrapers/collectlk.py — research collector used to build this snapshot. - scrapers/common.py — shared collector helpers. - scrapers/worldlib.py — shared fetch (live official URL, Wayback fallback). - scrapers/archivefallbacks.py — Wayback / Common Crawl CDX helpers. Official government /…
Research snapshot of official legislation collected from Journal Officiel (IORT/JORT; iort.gov.tn / ocr.jort.tn). Not legal advice. Official gazettes / government portals prevail over this corpus. - data/laws.parquet — one row per instrument (id, title, full text, dates, status, identifiers, article count, source URL). - data/articles.parquet — one row per article/section (lawid, article number, title, text, source URL). - scrapers/collecttn.py — research collector used to build this snapshot. - scrapers/common.py — shared collector helpers. - scrapers/worldlib.py — shared fetch (live official URL, Wayback fallback). - scrapers/archivefallbacks.py — Wayback / Common Crawl CDX helpers.…
Training data and model artifact dataset for the Aarhus weather pipeline. Maintained by Ciroc0. - targettimestamp - referencetime - leadtimehours - leadbucket - dmipred - actual - forecastwindu - forecastwindv - dmipredrundelta - observationcontexttimestamp - obs - tempcorrectiontarget - windspeedcorrectiontarget - windgustcorrectiontarget This dataset is released under CC BY 4.0. - Ciroc0 - Open-Meteo - https://open-meteo.com - DMI / DMI HARMONIE This dataset contains derived weather data and model artifacts built from upstream services. Attribution requirements from upstream providers still apply.
Model Collections
Hand-picked starting points, each with the reason it exists.
Collection · 4 entries
Embedding models for retrieval
Sentence and document embedding models used to build retrieval systems. Dimension and sequence length matter more than size here, and both come from the publisher.
Collection · 4 entries
Models that fit on one accelerator
Models whose publisher-reported parameter count puts them within reach of a single accelerator at common precisions. Memory needed depends on precision and serving configuration, so treat the parameter count as the starting point, not the answer.
Collection · 6 entries
Open-weight text models worth knowing
Widely used open-weight language models, chosen because each one is a distinct family rather than a variant of the one above it. Selection, not a ranking.
Collection · 3 entries
Speech and audio models
Recognition and synthesis models, grouped so the two directions are easy to compare.
Related SAVRN Research
The hub sits beside SAVRN's market data and infrastructure research: what models cost to run, and what it takes to run them.
SAVRN Index
What open models cost to run
The same open-weight model priced by every host that serves it, per million tokens.
Research Hub
Data center trackers and maps
Moratoriums, permits, power, water and capital behind the facilities that run these models.
Method
How the Model Hub is built
Sources, evidence labels, refresh behaviour, and the limits of every comparison here.