NOTICE

Tanor/SRGPTSENTNEG2
Model and documentation license: Creative Commons Attribution-ShareAlike 4.0 International (CC BY-SA 4.0)
License text: LICENSE

Original base model: jerteh/gpt2-orao
Upstream attribution: Mihailo Škorić; published by JeRTeh
Upstream model card: https://huggingface.co/jerteh/gpt2-orao/blob/a7dfec6b4290bcb48e579e13e39d20d0ea574449/README.md
Upstream license declaration: cc-by-sa-4.0
The upstream card revision above is the licensing-source snapshot inspected on
23 September 2026; it is not asserted to be the original training revision.

Base-model reference:
Škorić, Mihailo (2024). Novi jezički modeli za srpski jezik. Infoteka, 24(1). https://arxiv.org/abs/2402.14379

Fine-tuning and release: Tanor (Saša Petalinkar)
This model adapts the base-model family for binary NEG classification of
Serbian WordNet synset glosses using expansion iteration T2.
Documented fine-tuned weights revision: 609984070541e85cc2f51e0bafdfeeb42653c25c

Research reference:
Saša Petalinkar, Ranka M. Stanković, and Milica Ikonić Nešić (2025).
Comparative analysis of methods for creating a sentiment lexicon of the Serbian WordNet.
The Electronic Library, 43(4), 547–577.
https://doi.org/10.1108/EL-08-2024-0253
Companion code: https://github.com/sasa5linkar/Serbian-WordNet-Sentiment-Lexicon-Analysis

License documentation change, 23 September 2026:
The `cc-by-sa-4.0` declaration, full license text, and attribution were added on 23 September 2026.
Model weights, configuration, and tokenizer files were not modified.

External data and code keep their own terms. The exact SrpWN training-release
permissions remain to be documented; see README.md, License / Scope and
training-resource permissions. This notice provides attribution and provenance;
it does not add conditions to the license. Warranty and liability provisions
are set out in LICENSE.