A lora based on google/mt5-large a fine-tunes on indiejoseph/yue-zh-translation dataset. It would translate Mandarin to Cantonese. e.g.
input: translate Mandarin to Cantonese: 我都不知道你在说什么
output: 我都唔知你講咩
input: translate Mandarin to Cantonese: 整天就知道打游戏
output: 成日就知打遊戲
- Downloads last month
- 8
Inference API (serverless) does not yet support peft models for this pipeline type.