You can still use Prodigy for text classification. If you use a smaller vector model this issue with the serialization won’t occur, preventing the problem you were initially trying to work around. If necessary you can just not use the word vectors — they’ll improve accuracy, but it’s still definitely possible to train without them.
Related Topics
Topic | Replies | Views | Activity | |
---|---|---|---|---|
Large Datasets Google Cloud
|
5 | 1736 | October 13, 2018 | |
Bus Error/Segmentation Fault - Custom Gensim Vectors | 3 | 762 | July 10, 2018 | |
NER training - high RAM usage - Memory leak ? | 5 | 1197 | July 5, 2018 | |
biomedical nlp models in spacy | 4 | 2328 | February 28, 2018 | |
Memory Error | 3 | 2083 | June 15, 2018 |