Name | Modified | Size | Downloads / Week |
---|---|---|---|
Parent folder | |||
README.md | 2021-10-27 | 480 Bytes | |
Updated multilingual model _ consistent class names source code.tar.gz | 2021-10-27 | 234.7 kB | |
Updated multilingual model _ consistent class names source code.zip | 2021-10-27 | 248.3 kB | |
Totals: 3 Items | 483.5 kB | 0 |
- Updated the multilingual model weights used by Detoxify with a model trained on the translated data from the 2nd Jigsaw challenge (as well as the 1st). This model has also been trained to minimise bias and now returns the same categories as the unbiased model. New best AUC score on the test set: 92.11 (89.71 before).
- All detoxify models now return consistent class names (e.g. "identity_attack" replaces "identity_hate" in the original model to match the unbiased classes).