RedReddington1's picture
Update README.md
b2fbe01
metadata
language:
  - en
tags:
  - pytorch
  - causal-lm
  - pythia
license: apache-2.0
datasets:
  - bigbio/med_qa

This adapter was lora-trained using lit-gpt for 12,000 steps (approx. 1 epoch) using Pythia-410M-deduped model on Med_qa english dataset.

Prompt template:

Below is an Question that is a medical problem, paired with an input that provides the answer choices. Write a response that accurately answers the question.

### Question:
[QUESTION]

### Options:
[OPTIONS]

### Response:

The options is a string like "A. Option 1\nB. Option B\nC. Option C"