Open sambalshikhar opened 3 weeks ago
The current checkpoints are pre-trained or fine-tuned on vision-language tasks and are not suitable for language-specific tasks. I will upload checkpoints trained on open-source datasets, such as summarization tasks based on HealthCare-Magic. However, for datasets derived from sources like MIMIC-III, I am unable to distribute the checkpoints due to licensing restrictions, which require permission for distribution. If you are interested in reproducing the results, you can fine-tune the models yourself. Alternatively, please email me directly at kaz321@lehigh.edu to get the checkpoints, including the PhysioNet license, ensuring it aligns with the email you send.
Thank you for your interest in this work.
Hi , i am trying to evaluate the natural language task on mednli and summerisation. However the inference gives out results that do not make sense.I guess the checkpoints for natural language fine tuning has not been provided.