README.md

Fine-tuned Model Submission Template

This is a template reprository for the SUPERB benchmark for the fine-tuned model category. In this category, participants are asked to fine-tuned a pretrained model in each of SUPERB's downstream tasks and then store the model weights and hyperparameters in this repo.

There are four steps involved in making a submission:

1. Fine-tune a pretrained model on a downstream task.
2. Implement the PreTrainedModel interface defined in each model.py module.
3. Store the weights and hyperparameters in the task directory
4. Push all the files to the Hugging Face Hub.