huseinzol05/t5-small-bahasa-cased

The model was trained using Google T5's Github repository on v3-8 TPU . It was used to train T5 small language models for Malay and Indonesian . The model has been pretrained on multiple tasks . It can be pre-processed using torch or tensorflow and Huggingface library transformers . The results can be reproduced from here, Malaya/pretrained-model/t5.io/en-latest/Accuracy . The full training steps can be found here,  at Malaya.org/malaya/Pretrained/pretraining-model.html, we compared with traditional models. For further details on the model performance,