-
Notifications
You must be signed in to change notification settings - Fork 13
Fine Tuning a Model on a Custom Dataset
Hugging Face is a platform that enables users to build build, train and deploy ML models based on open source code and technologies. It has a lot of pre-trained models for specific tasks that save the time and computation required to train large models.
Hugging Face has been used for the fine-tuning task because of the large number of pre-trained models already available. We can use the t5-large model that has already been trained on an English to Spanish machine translation task. Furthermore, HF also has the ability to upload trained models which can be used by others easily. They have an API interface which can do the required task right from the website and therefore, because of these multiple reasons Hugging Face was used to fine-tune the model.