Skip to content

Latest commit

 

History

History
61 lines (47 loc) · 1.31 KB

README.md

File metadata and controls

61 lines (47 loc) · 1.31 KB

In-progress

Overview

Fine tune gpt3 model, and create news prompt summary.

Getting Started

dataset for training:

kaggle datasets download -d sunnysai12345/news-summary
python format_dataset.py

This creates "dataset.jsonl"

openai tools fine_tunes.prepare_data -f dataset.jsonl

this command will create a file "dataset_prepared.jsonl"

openai api fine_tunes.create -t "dataset_prepared.jsonl" -m ada

this command will provide a model name

Usage

python run_model.py --fine_tuned_model <YOUR_MODEL_NAME>

or

openai api completions.create -m <YOUR_MODEL_NAME> -p <YOUR_PROMPT>

Example Prompt

Technology incubator Jigsaw, a subsidiary of Google-parent Alphabet, has created an online quiz to help test users' ability to detect fake emails meant for 'phishing' attacks to steal passwords or download malware. The quiz tests and teaches users to judge factors like email address to figure out the authenticity of the emails. Jigsaw further suggests users enable '2-step Verification' feature.
Summary:

Output

{
  "choices": [
    {
      "finish_reason": "length",
      "index": 0,
      "logprobs": null,
      "text": " Alphabet govt tech incubator creates quiz to help test fake emails\n\nAl"
    }
  ],
  ...
}