Fine tuned the open source seq2seq models

[Chetan-Yeola]

Hello Team,
Can I use Colossalai to fine-tune the Text-to-Text generation model on a custom dataset? If so, please describe the procedures.

[binmakeswell]

Hi @Chetan-Yeola Yes, it accelerates training speed and reduces hardware resource requirements without limiting your model and dataset. You can refer to our examples:
https://github.com/hpcaitech/ColossalAI/tree/main/examples
https://github.com/hpcaitech/ColossalAI/tree/main/applications/Chat
If you need customized in-depth cooperation or support, please send the details to [email protected]