SETimes_sr is a Serbian dataset annotated for morphosyntactic information and named entities.
The dataset contains 3177 training samples, 395 validation samples and 319 test samples
across the respective data splits. Each sample represents a sentence and includes the following features:
sentence ID ('sent_id'), sentence text ('text'), list of tokens ('tokens'), list of lemmas ('lemmas'),
list of Multext-East tags ('xpos_tags), list of UPOS tags ('upos_tags'),
list of morphological features ('feats'), and list of IOB tags ('iob_tags'). The 'upos_tags' and 'iob_tags' features
are encoded as class labels.
