Finetuned Models

[jungokasai]

It looks that the shared models are pretrained models before finetuning. Would it be possible to get access to two finetuned models (MIMIC-CXR/Open-I generation finetuned models)? I would use them exclusively for research purposes and cite this work accordingly. Thank you so much for this repository.

Jungo https://homes.cs.washington.edu/~jkasai/

[SuperSupermoon]

@jungokasai I'm sorry that I can't directly provide the dozens of finetuned models (56 different models: 7 models in each task x 4 downstream tasks x 2 datasets). But it is rather straightforward. Using the pre-trained models we provide, you can fine-tune the model for each downstream task.