Kashgari is a production-level NLP Transfer learning framework built on top of tf.keras for text-labeling and text-classification, includes Word2Vec, BERT, and GPT2 Language Embedding.

中文长文本分类、短句子分类、多标签分类、两句子相似度（Chinese Text Classification of Keras NLP, multi-label classify, or sentence classify, long or short），字词句向量嵌入层（embeddings）和网络层（graph）构建基类，FastText，TextCNN，CharCNN，TextRNN, RCNN, DCNN, DPCNN, VDCNN, CRNN, Bert, Xlnet, Albert, Attention, DeepMoji, HAN, 胶囊网络-CapsuleNet, Transformer-encode, Seq2seq, SWEM, LEAM, TextGCN

自然语言处理（nlp），小姜机器人（闲聊检索式chatbot），BERT句向量-相似度（Sentence Similarity），XLNET句向量-相似度（text xlnet embedding），文本分类（Text classification）， 实体提取（ner，bert+bilstm+crf），数据增强（text augment, data enhance），同义句同义词生成，句子主干提取（mainpart），中文汉语短文本相似度，文本特征工程，keras-http-service调用

Simple State-of-the-Art BERT-Based Sentence Classification with Keras / TensorFlow 2. Built with HuggingFace's Transformers.

Bert-classification and bert-dssm implementation with keras.

本项目采用Keras和Keras-bert实现文本多分类任务，对BERT进行微调。

Deep learning models(CNN, LSTM, BERT) for image and text classification task with Tensorflow and Keras

Fine-tuning google's BERT model on text classification for sentiment analysis with pytorch and keras

fine-tuned BERT and scikit-learn models for real-time classification of disaster-related tweets, using TensorFlow, Keras, and Transformers. .

In this project textual as well as visual features are used to classify image documents. OCR is implemented to get text data from image documents then BERT model is used for text embedding. After this Hierarchical Attention Network HAN with word and sentence attention along with GRU (Gated Recurrent Unit ) is used to classify documents. Second we have done simple image classification on our documents using only visual features. After training both models individually on training data we checked our models performance on test set. We got vectors from both models, combine them and used this vector to classify our documents.The idea is taken from currently published research paper known as "Efficient multi ti modal Document Image Classifier for Scarce Data" (January,2022) at https://www.mdpi.com/2076-3417/12/3/1457. Pytorch and Keras are the main deep learning frameworks that are used in this project.

No description available

Majority of the commercial music platforms rely heavily on deep learning and natural language processing for the purpose of finding similarities between songs, classification of songs, new lyric generation, informational retrieval, gaining meaningful insights or efficiently analyze music. My main focus was to use raw lyrics data and classify them as one of the 5 genres (Hip-Hop, Rock, Pop, Country and Jazz). For this purpose, architecture such as Machine learning classifiers, Deep Neural Networks, LSTM and Bi-Directional Transformers were investigated. Various Data preprocessing and network optimization techniques were utilised and While, It was observed that keras BERT model and LSTM produced similar results, BERT worked the best than the rest.

No description available

No description available

BERT Classification of Reddit AITA Question Threads (BERT, NLP, Deep Learning, Keras, Tensorflow, CUDA)

Sentiment analysis on IMDB movie reviews using Python. Compares LSTM, CNN, and BERT models for binary classification. Includes preprocessing, model training, evaluation, and performance comparison. Built with TensorFlow/Keras and Hugging Face.

Implementaion of two task related to Binary Classification and POS TAGGING. Use of Transformer pretrained model (BERT-uncase) and some techniques (freezing etc) for optimizing the tasks. Use both pytorch and keras for different implemenation.

No description available

No description available

通過 Keras 微調 Bert 模型，可對文本或代碼進行標籤與分類

No description available

It took me a month to come to a comprehensible and tidy model of Bert for token classification using keras; there were numerous versions which were not at all comprehensible to me as a beginner in Bert. Accordingly I decided to upload it to my github so that it may be helpful to you to learn Bert from scratch,

A guide to mastering text classification with BERT and KerasNLP

Text classification using Bert, Tensorflow and Keras.

NLP keras_bert model for binary classification.

Multiclass text classification using two models : Keras Custom Embeddings and Google BERT

fine tuned BERT for binary classification of food-related tweets as healthy/unhealthy, using TensorFlow, Keras, and Transformers.

Classifying products into predefined categories (supervised learning multi-class classification) using keras tensorflow and a pre-trained transformer-based BERT model.

BERT-based transformer with soft-voting ensemble and per-label threshold optimization to handle class imbalance in multi-emotion text classification using TensorFlow/Keras.

a text classification project featuring two deep learning approaches: an LSTM-based model built with TensorFlow/Keras and a BERT-based model fine-tuned with Hugging Face Transformers. Both notebooks include end-to-end text preprocessing, tokenization, model training, and evaluation. Ask ChatGPT