This repo illustrates how to use OpenAI GPT to
- generate a synopsis from a podcast transcription into multiple languages.
- generate 2-3 tag lines based on the podcast content.
- generate serch engine optimised (SEO) keywords.
This example uses the transcription of podcast : Why SVB’s collapse is not a 2008 repeat avaiable from Financial Times to illustrate how to make use of OpenAI GPT for the use case described above.
- podcast-to-synopsis-taglines-keywords.ipynb This notebook explores how to split a long podcast transcript into chuncks, generate synopsis, taglines, and search engine optimised (SEO) keywords.
- podcast-to-synopsis-various-genres.ipynb This notebook explores how prompt design can influence the output of GPT in the context of podcast synopsis generation.
- Transcribe from audio to text.
- Cater for large data corpus.
Related repos that applied this concept:
Related blog post that applied this concept:
Related repos from the author:
- https://github.com/ryubidragonfire/document-analysis-using-gpt-3
- https://github.com/ryubidragonfire/code-explaination-openai
- https://github.com/ryubidragonfire/personalised-movie-reviews-gpt-3
- OpenAI repo: https://github.com/openai/openai-cookbook/blob/main/examples/Entity_extraction_for_long_documents.ipynb
- Which embedding model to use? https://openai.com/blog/new-and-improved-embedding-model/
- OpenAI models: https://learn.microsoft.com/en-us/azure/cognitive-services/openai/concepts/models
- API reference: https://platform.openai.com/docs/api-reference/completions/create