rostlabs/rost-286m-sft

286M parameter nanochat GPT (sft stage), research tag normalized-d12-004b, checkpoint step 000600.

Part of the rost model zoo — the full set of trained variants behind the rost research series, published for reproducibility. This is a research model, not a product.

architecture	nanochat GPT, 12 layers, 768 embed, 2048 context
tokenizer	included under tokenizer/ (32,768 vocab)
training mixture	Romanian SFT mixture (OpenLLM-Ro datasets) with input diacritic augmentation
stage	sft
val bpb (own split)	0.49526

The best Romanian conversational model of the rost small series (research tag normalized-d12-004b). Input-side diacritic augmentation makes it robust to unaccented Romanian as actually typed (cum te cheama?). Base: rost-286m-base.

Validation bits-per-byte is measured on this arm's own validation split and is not comparable across arms — cross-arm comparisons in the rost write-ups are always cross-evaluated on identical text.

Licence

CC-BY-NC-4.0, non-commercial, inherited from the most restrictive component of the training data.