NER

46 public types in this namespace, organized by kind.

Models & Types (34)

Type Summary
BERTNER<T> BERT-NER: BERT (Bidirectional Encoder Representations from Transformers) with token classification head for Named Entity Recognition.
BLINKNER<T> BLINK: BERT-based bi-encoder for entity linking and Named Entity Recognition.
BiLSTMCRF<T> BiLSTM-CRF: Bidirectional LSTM with Conditional Random Field for Named Entity Recognition.
BiaffineNER<T> Biaffine-NER: Named Entity Recognition as dependency parsing using biaffine classifiers.
BioBERTNER<T> BioBERT-NER: Biomedical domain-specific BERT for Named Entity Recognition in biomedical text.
CNNBiLSTMCRF<T> CNN-BiLSTM-CRF: Character CNN + Bidirectional LSTM + Conditional Random Field for Named Entity Recognition.
ClinicalBERTNER<T> ClinicalBERT-NER: Clinical domain BERT for Named Entity Recognition in clinical notes and EHRs.
DeBERTaNER<T> DeBERTa-NER: Decoding-enhanced BERT with disentangled Attention for NER.
DistilBERTNER<T> DistilBERT-NER: Knowledge-distilled BERT for efficient Named Entity Recognition.
ELECTRANER<T> ELECTRA-NER: Efficiently Learning an Encoder that Classifies Token Replacements Accurately for NER.
FinBERTNER<T> FinBERT-NER: Financial domain BERT for Named Entity Recognition in financial text.
InstructionNER<T> InstructionNER: Instruction-tuned transformer for few-shot and zero-shot NER via natural language instructions.
LSTMCRF<T> LSTM-CRF: Unidirectional LSTM with Conditional Random Field for Named Entity Recognition.
LegalBERTNER<T> Legal-BERT-NER: Legal domain BERT for Named Entity Recognition in legal documents.
NERTrainingProgress Reports training progress for NER models, including loss, F1 score, and epoch information.
NamedEntitySpan An entity a span-based NER model found: tokens Start to End inclusive, of category Label, with the probability the model gave it.
NerTextEncoder Converts raw text into the packed integer indices consumed by WordCharEmbeddingLayer / WordCharBiLSTMCRF, and builds the word and character vocabularies.
ONNXNER<T> ONNX-NER: Generic ONNX Runtime-based NER model for high-performance inference with any exported NER model.
PURENER<T> PURE: Princeton University Relation Extraction - pipeline approach for joint entity and relation extraction.
PromptNER<T> PromptNER: Prompt-based learning for few-shot Named Entity Recognition.
PubMedBERTNER<T> PubMedBERT-NER: PubMed domain-specific BERT pre-trained from scratch on biomedical text for NER.
PyramidNER<T> Pyramid-NER: Hierarchical pyramid network for nested Named Entity Recognition.
RELNER<T> REL: Radboud Entity Linker - end-to-end entity linking combining NER with entity disambiguation.
RoBERTaNER<T> RoBERTa-NER: Robustly Optimized BERT Approach with token classification for NER.
SECBertNER<T> SEC-BERT-NER: Securities and Exchange Commission domain BERT for NER in regulatory filings.
SciBERTNER<T> SciBERT-NER: Scientific domain BERT for Named Entity Recognition in scientific literature.
SpERTNER<T> SpERT: Span-based Entity and Relation Transformer for joint entity and relation extraction.
SpanBERTNER<T> SpanBERT-NER: Span-level BERT pre-training with token classification for NER.
TemplateNER<T> Template-NER: Template-based prompt approach for few-shot and zero-shot NER.
TinyBERTNER<T> TinyBERT-NER: Two-stage distilled BERT for ultra-efficient Named Entity Recognition.
TriaffineNER<T> Triaffine-NER: Three-way interaction model for nested Named Entity Recognition.
W2NER<T> W2NER: Word-Word Relation Classification for unified flat and nested NER.
WordCharBiLSTMCRF<T> Paper-faithful word + character BiLSTM-CRF for Named Entity Recognition (Lample et al., NAACL 2016, “Neural Architectures for Named Entity Recognition”).
XLMRoBERTaNER<T> XLM-RoBERTa-NER: Cross-lingual RoBERTa for multilingual Named Entity Recognition.

Base Classes (4)

Type Summary
NERNeuralNetworkBase<T> Base class for NER-focused neural networks that can operate in both ONNX inference and native training modes.
SequenceLabelingNERBase<T> Base class for sequence labeling NER models that assign a BIO label to each token in a sequence.
SpanBasedNERBase<T> Base class for span-based NER models (SpERT, BiaffineNER, PURE).
TransformerNERBase<T> Base class for transformer-based NER models (BERT-NER, RoBERTa-NER, DeBERTa-NER, etc.).

Interfaces (1)

Type Summary
INERModel<T> Base interface for all Named Entity Recognition (NER) AI models in AiDotNet.

Options & Configuration (7)

Type Summary
BiLSTMCRFOptions Configuration options for the BiLSTM-CRF Named Entity Recognition model.
BiaffineNEROptions Options for Biaffine-NER, defaulting to the values published in Yu et al., ACL 2020, “Named Entity Recognition as Dependency Parsing” (arXiv:2005.07150).
CNNBiLSTMCRFOptions Configuration options for the CNN-BiLSTM-CRF Named Entity Recognition model.
LSTMCRFOptions Configuration options for the LSTM-CRF Named Entity Recognition model.
PromptNEROptions Configuration for PromptNER’s paper architecture and training recipe.
SpanBasedNEROptions Options shared by span-based NER models (SpERT, BiaffineNER, PURE).
TransformerNEROptions Base configuration options shared by all transformer-based NER models (BERT-NER, RoBERTa-NER, DeBERTa-NER, ELECTRA-NER, etc.).