A small T5v1.1 model trained on the cLang-8 dataset following the process described in A Simple Recipe for Multilingual Grammatical Error Correction.
Since it is a single-task T5v1.1 model, there is no need for a task-specific prefix.
- Downloads last month
- 70