The data here seems good enough on its own to train a pretty strong NNUE. In my own tests it did a good job of pulling draws out of Stockfish:P To keep costs down, it was relabeled with t1-256x10-rl-base-swa-3860000, and most of the self-play games were also generated with that same model, in a single pass with a bit of temperature. Bigger lc0 models would be far more effective, but this one hit around 660k/s on 8× 5090s, which made it cheap, so that's what I went with.
Independent publisher
Çetin
cturan
Models in Library0
Datasets in Library1
Models on Hugging Face6
Followers12
Compare the published evidence
This profile connects the 1 model and dataset record shown below to their upstream cards and sources. The library counts describe SAVRN's coverage, rather than this publisher's entire output.
Reported licenses: odc-by. Open a record to review its specific license and conditions.
Compare model specifications, source dates, and hardware estimates on the individual records. Publisher identity verification is separate from a model's quality, deployment readiness, or suitability for your work.