gpt2-large-medical/README.md

datasets, metrics, pipeline_tag, widget, inference, tags

datasets	metrics	pipeline_tag	widget	inference	tags
BI55/MedText pubmed_qa	bleu perplexity	text-generation	text <|USER|> What is Schizophrenia? <|ASSISTANT|>	parameters temperature top_p top_k max_new_tokens repetition_penalty 0.8 0.14 41 15 1.176	medical

A further fine-tuned version of Locutusque/gpt2-large-conversational on MedText and pubmed_qa

Evaluation

This model was evaluated using GPT-3.5, and it was asked medical questions. It achieved an average accuracy of 80%.