Frequency-Agnostic
Word embedding trainer
Improves word embeddings by using adversarial training to make them less dependent on word frequencies
Code for NIPS 2018 paper 'Frequency-Agnostic Word Representation'
118 stars
6 watching
20 forks
Language: Python
last commit: over 5 years ago Related projects:
| Repository | Description | Stars |
|---|---|---|
 hit-scir/elmoformanylangs | Provides pre-trained ELMo representations for multiple languages to improve NLP tasks. | 1,463 |
 jwieting/paragram-word | Trains word embeddings from a paraphrase database to represent semantic relationships between words. | 30 |
| jwieting/charagram | A tool for training and using character n-gram based word and sentence embeddings in natural language processing. | 125 |
 wangyuxinwhy/uniem | Develops unified sentence embedding models for NLP tasks | 833 |
 vzhong/embeddings | Provides fast and efficient word embeddings for natural language processing. | 223 |
 zhangxiann/skip-gram | A Python implementation of a neural network model for learning word embeddings from text data | 6 |
 malllabiisc/wordgcn | A deep learning model that generates word embeddings by predicting words based on their dependency context | 290 |
 kyubyong/wordvectors | Provides pre-trained word vectors for multiple languages to facilitate NLP tasks | 2,215 |
 seomoz/word2gauss | This implementation provides a way to represent words as multivariate Gaussian distributions, allowing scalable word embeddings. | 190 |
 fanglanting/skip-gram-pytorch | A PyTorch implementation of the skip-gram model for learning word embeddings. | 188 |
 hassygo/charngram2vec | A repository providing a re-implementation of character n-gram embeddings for pre-training in natural language processing tasks | 23 |
 harsh19/spine | Transforms existing word embeddings into more interpretable ones by applying a novel extension of k-sparse autoencoder with stricter sparsity constraints | 52 |
 ray1007/gwe | A software implementation of a word embedding method using character glyphs, enhancing traditional Chinese language processing | 30 |
 zhezhaoa/ngram2vec | A toolkit for learning high-quality word and text representations from ngram co-occurrence statistics | 846 |
 nikhgarg/embeddingdynamicstereotypes | Quantifies 100 years of gender and ethnic stereotypes through word embeddings. | 159 |