18 packages tagged with “BERT”
Fast and memory-efficient WordPiece tokenizer as it is used by BERT and others. Tokenizes text for further processing using NLP/language models.
This package contains tokenizers for following models: · BERT Base · BERT Large · BERT German · BERT Multilingual · BERT Base Uncased · BERT Large Uncased
NET Standard 2.1 library to produces embeddings using C# Bert Tokenizer and Onnx All-Mini-LM-L6-v2 model.
BERT serializer and BERT-RPC client for .Net
Text processing infrastructure for LMSupply - tokenization, vocabulary management, and encoding utilities
Minimal Tokenizer implementation of BertJapanese(tohoku-nlp/bert-base-japanese) in C#
Erlang External Term Format (ETF/BERT) encoder/decoder for F#
Package Description
Text translator library based on LLM models, especially EncoderDecoderModel in HuggingFace
Local-first document summarization library using BERT embeddings, RAG, and optional LLM synthesis. Supports markdown, PDF, DOCX, and URLs. Every claim is grounded with citations. Runs entirely offline with ONNX models, or optionally uses Ollama/Docling for enhanced features.
A simple .NET library for local text embeddings with automatic model downloading from HuggingFace. Supports CUDA, DirectML, and CoreML GPU acceleration.
A lightweight, zero-configuration semantic reranker for .NET. Supports multiple cross-encoder models with automatic download, GPU acceleration, and HuggingFace caching.
A BERT tokenization extension for EasyReasy KnowledgeBase with FastBertTokenizer integration
A .NET Framework 4.6.1 compatible implementation of Sentence Transformers in C#. Produces embeddings using C# BERT Tokenizer and ONNX All-Mini-LM-L6-v2 model. Includes ONNX Runtime initialization helpers for legacy .NET Framework environments. Perfect for semantic search, text similarity, and embedding generation in .NET applications.
Wordpiece tokenizer for using eg onnx models with ML.Net
High-performance WordPiece/BERT tokenizer in C# (port of FlashTokenizer)
Local-first OCR, Named Entity Recognition, and Vision captioning library. Uses Tesseract OCR, BERT NER (ONNX), Florence-2 vision, and ImageSharp preprocessing. Auto-downloads all models on first use - zero manual setup required.