early-stopping

问题 I am training a skipgram model using gensim word2vec. I would like to exit the training before reaching the number of epochs passed in the parameters based on a specific accuracy test in a different set of data in order to avoid the overfitting of the model. Is there a way in gensim to interrupt the train of word2vec from a callback function? 回答1: If in fact more training makes your Word2Vec model worse on some external evaluation, there is likely something else wrong with your setup. (For

问题 I am training a skipgram model using gensim word2vec. I would like to exit the training before reaching the number of epochs passed in the parameters based on a specific accuracy test in a different set of data in order to avoid the overfitting of the model. Is there a way in gensim to interrupt the train of word2vec from a callback function? 回答1: If in fact more training makes your Word2Vec model worse on some external evaluation, there is likely something else wrong with your setup. (For