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.
Updated 2026-10-18 · How the library is built
3,164 datasets, sorted by most downloaded.
The runs write here while they train; each run has its own folder (its train.hubprefix). filters, the 236 held-out and 177 echo-unseen voices left out. Nothing else: the run trains only on the provided data. in the screening runs of VoiceHub/mytts-en-ablations; REPA alignment to w2v-BERT 2.0 layer 16. Configuration: configs/train/enfull.yaml in kadirnar/dacvae-next (branch roadmap/en-echo; private at the moment). apache-2.0 (owner decision, 2026-09-29). Trained only on SynDataLab-EN/echo-clones-4m-en (apache-2.0 on its card), synthetic speech generated with EchoTTS (jordand/echo-tts-base, cc-by-nc-sa-4.0 on its card). The provenance notes are on the model page.
This repository contains the current selected versions and owner-feedback revisions replacing the earlier 2,500-video campaign. The old campaign videos and their metadata have been removed from the current dataset. The repository remains private. The feedback revision scope is 77 families with five samples each: 385 / 385 revised clips are available. An additional 115 / 115 clips cover the 23 families outside that feedback list, using one selected version and five samples per family. The other-session selections are preserved except G-174, whose five clips were explicitly replaced by the newly requested kitchen revision. New completed revisions will be added incrementally; this is not the…
Benchmark for Evaluating Tool-Augmented PII Detection in Small Language Models, a technical report (not peer reviewed). The 2,000 benchmark rows are unchanged. All canonical label counts agree with the primary scoring gold after the report's normalization. The benchmark contains labels and source indices; it does not contain a complete span-annotated dataset. Earlier descriptions of this release as uniformly span-annotated were inaccurate. There are 1,132 AI4Privacy, 581 Gretel and 287 NVIDIA examples. The scoring gold used for the historical report has spans for the AI4Privacy subset and type fallback for the other 868 examples. NVIDIA source annotations do have offsets; the historical…
What came before the Salinas City Council and its committees, what was done with it, and how each member voted. salinas.legistar.com, read through its public API at webapi.legistar.com/v1/salinas. Nothing is added to what the City entered there, and nothing is corrected. Agendas here run from 2014-03-25 to 2026-10-15, across 21 bodies: the Council, its committees, and the boards and commissions the City keeps in the same system. Votes are recorded from 2014-04-08 to 2026-09-29. The five tables, and the columns listed first in each, are the same in every council and board dataset made this way, so two can be set one on the other: and are two. matterid joins items to agendaitems and votes…
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.