README.md

Setup

1. Get you API KEY form openai.com
2. Create a file called .env and put your API KEY in it

OPENAI_API_KEY=XYZ

3. Install venthur/dotenv-cli

   pip install dotenv-cli

4. Install Poetry and run poetry install

Usage

1. Single question:

   dotenv --dotenv .env poetry run askgpt "Why Amsterdam is known of diamonds?"

2. Single question using docker:

   docker run -it --env-file .env askgpt "Why Amsterdam is known of diamonds?"

3. Summarize long text

   cat article5.txt | dotenv --dotenv .env poetry run askgpt --summarize --max-tokens 256

4. Get plaintext from URL:

   docker run -it trafilatura -u "https://roadtoomega.substack.com/p/the-paradigm-of-emergence-a-unifying" > article5.txt

5. Prefix article with question:

   echo -e "Write a concise summary of the following text:\n" | cat - article5.txt | dotenv --dotenv .env poetry run askgpt --max-tokens 500

References

Loosely related to the topic

1. openai-cookbook/techniques_to_improve_reliability.md at main · openai/openai-cookbook
2. Summarizing Books with Human Feedback
3. Learning to Summarize with Human Feedback
4. bart-large-cnn
   • huggingface/transformers: 🤗 Transformers: State-of-the-art Machine Learning for Pytorch, TensorFlow, and JAX.
   • fairseq/examples/bart at main · facebookresearch/fairseq
5. daveshap/RecursiveSummarizer