mesolitica/nanot5-base-malaysian-translation-v2.1

NanoT5 Base Malaysian Translation v2.1

Finetuned https://huggingface.co/mesolitica/nanot5-base-malaysian-cased using 2048 context length on 9B tokens of translation dataset.

• —This model able to translate from localize text into standard text.
• —This model able to reverse translate from standard to localize text, suitable for text augmentation.
• —This model able to translate code.
• —This model natively code switching.
• —This model should maintain \n, \t, \r as it is.
• —Better Science and Math context translation compared to v2.
• —Better Manglish translation compared to v2.
• —Better Cantonese translation compared to v2.
• —Better Tamil and Tanglish translation compared to v2.

Wandb at https://wandb.ai/huseinzol05/nanot5-base-malaysian-cased-translation-v6-multipack-post

how to

You can follow the same how to from the small model README, https://huggingface.co/mesolitica/nanot5-small-malaysian-translation-v2.1#how-to