indonesian-roberta-base-posp-tagger
indonesian-roberta-base-posp-tagger performs sequence labeling: each input token receives a class label aligned to its text position. Typical tasks include NER, chunking, and slot filling.
2,780,091 ↓ · 10 ♡
token classification models
20 models · ranked by HuggingFace downloads
bert-base-NER
bert-base-NER performs sequence labeling: each input token receives a class label aligned to its text position. Typical tasks include NER, chunking, and slot filling.
1,687,260 ↓ · 718 ♡
stanford-deidentifier-base
stanford-deidentifier-base uses a BERT encoder with a per-token classification head. The BIO tagging scheme is standard for its NER fine-tunes.
1,220,506 ↓ · 81 ♡
wikineural-multilingual-ner
wikineural-multilingual-ner performs sequence labeling: each input token receives a class label aligned to its text position. Typical tasks include NER, chunking, and slot filling.
823,591 ↓ · 165 ♡
finetuned-bertweet-poskd3
A DistilBERT model fine-tuned on Twitter/X data for token classification tasks, likely part-of-speech tagging or named entity recognition on social media text. BERTweet-based initialization means it handles informal spelling, hashtags, and abbreviations better than standard BERT. The training split and label schema are not publicly documented.
749,718 ↓ · 0 ♡
fullstop-punctuation-multilang-large
fullstop-punctuation-multilang-large assigns labels to individual tokens in a sequence, directly applicable to named entity recognition, part-of-speech tagging, and span extraction.
667,216 ↓ · 177 ♡
bert-large-cased-finetuned-conll03-english
bert-large-cased-finetuned-conll03-english assigns labels to individual tokens in a sequence, directly applicable to named entity recognition, part-of-speech tagging, and span extraction.
660,324 ↓ · 96 ♡
punctuate-all
punctuate-all uses a RoBERTa encoder with a per-token classification head. The BIO tagging scheme is standard for its NER fine-tunes.
525,030 ↓ · 28 ♡
xlm-roberta-large-ner-hrl
xlm-roberta-large-ner-hrl uses a RoBERTa encoder with a per-token classification head. The BIO tagging scheme is standard for its NER fine-tunes.
524,801 ↓ · 15 ♡
layoutreader
layoutreader is an open-source token-classification model available on HuggingFace. Details are sourced from the public model registry.
508,613 ↓ · 43 ♡
sat-3l-sm
SAT-3l-sm (Segment Any Text, 3-layer small) is a multilingual text segmentation model supporting over 85 languages, designed to split continuous text into meaningful sentence or paragraph segments. Unlike rule-based sentence tokenisers that rely on punctuation, SAT uses a contextual XLM-based token classifier to handle languages with unusual or absent punctuation conventions. The small variant trades some accuracy for faster inference.
459,079 ↓ · 12 ♡
deid_roberta_i2b2
deid_roberta_i2b2 is a RoBERTa model fine-tuned on the i2b2 de-identification dataset to detect and classify protected health information (PHI) in clinical notes. It identifies PHI spans such as names, dates, locations, and IDs. MIT-licensed for integration into clinical NLP de-identification pipelines.
431,180 ↓ · 39 ♡
llmlingua-2-xlm-roberta-large-meetingbank
llmlingua-2-xlm-roberta-large-meetingbank assigns labels to individual tokens in a sequence, directly applicable to named entity recognition, part-of-speech tagging, and span extraction.
430,411 ↓ · 28 ♡
ner-english-fast
Flair's fast English NER model using the Flair framework's sequence labeling approach with character-level language model embeddings. 'Fast' indicates a smaller, speed-optimized variant compared to Flair's standard NER model. Recognizes standard NE classes (PER, ORG, LOC, MISC).
422,834 ↓ · 26 ♡
bert-base-NER-Russian
A BERT-base model fine-tuned for named entity recognition on Russian text. Handles standard NER categories (persons, organizations, locations) on Russian-language inputs using the standard token-classification approach.
357,936 ↓ · 22 ♡
privacy-filter
privacy-filter is an open-source token-classification model available on HuggingFace. Details are sourced from the public model registry.
316,092 ↓ · 1,583 ♡
roberta-large-ner-english
roberta-large-ner-english is an open-source token-classification model available on HuggingFace. Details are sourced from the public model registry.
308,637 ↓ · 79 ♡
bert-portuguese-ner
bert-portuguese-ner is an open-source token-classification model available on HuggingFace. Details are sourced from the public model registry.
304,136 ↓ · 9 ♡
bert-base-chinese-ws
bert-base-chinese-ws is an open-source token-classification model available on HuggingFace. Details are sourced from the public model registry.
296,571 ↓ · 19 ♡
bert-base-multilingual-cased-ner-hrl
bert-base-multilingual-cased-ner-hrl is an open-source token-classification model available on HuggingFace. Details are sourced from the public model registry.
292,339 ↓ · 82 ♡