Conversation Data

[ChantalMP]

Hi,
great work! thanks for publishing you code and data.
What exactly do the radiology conversations in Healthcare_Radiology_vicuna.json contain? It still seems to be related to the data from ChatDoctor - did you somehow filter this data or how did you generate the radiology conversation data?

Thanks!

[Amshaker]

Hi @ChantalMP ,

Thank you for your interest in our work and sorry for the late reply.

We fine-tuned the original Vicuna using two stages. The first stage is using medical data from ChatDoctor, just removing ChatDcotor keywords and aligning it in the format of Vicuna. In the second fine-tuning stage, we use radiology-related search terms to filter out the data and obtain a 22k subset for fine-tuning.

I hope it is clear now.

Best regards,
Abdelrahman.