How to feed data for completions, instead of using prompt/answer fine-tuning format?

ishikawa · March 8, 2023, 2:11pm

Very good. Did you find some tutorial about covering all this process?

chirag.shah285 · March 8, 2023, 2:41pm

upvoting this! Any tutorials? Did you find some tutorial about covering all this process?

jiangweidong · March 15, 2023, 3:16am

You can ref this notebook to use embedding to answer questions based on own knowlage.

openai/openai-cookbook/blob/c3b920f123a9c92c30b106462b5d5ecf53a4b3ab/examples/Question_answering_using_embeddings.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "c4ca8276-e829-4cff-8905-47534e4b4d4e",
   "metadata": {},
   "source": [
    "# Question Answering using Embeddings\n",
    "\n",
    "Many use cases require GPT-3 to respond to user questions with insightful answers. For example, a customer support chatbot may need to provide answers to common questions. The GPT models have picked up a lot of general knowledge in training, but we often need to ingest and use a large library of more specific information.\n",
    "\n",
    "In this notebook we will demonstrate a method for enabling GPT-3 to answer questions using a library of text as a reference, by using document embeddings and retrieval. We'll be using a dataset of Wikipedia articles about the 2020 Summer Olympic Games. Please see [this notebook](fine-tuned_qa/olympics-1-collect-data.ipynb) to follow the data gathering process."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "9e3839a6-9146-4f60-b74b-19abbc24278d",
   "metadata": {},
   "outputs": [],

This file has been truncated. show original