Automata Learning Lab
In this video, I will quickly go over my medium article showing how to use OpenAi’s API to interact with the GPT-3 model for summarizing a machine learning paper page by page.
– Link for my article on Medium: https://medium.com/geekculture/a-paper-summarizer-with-python-and-gpt-3-2c718bc3bc88
– Join Medium: https://lucas-soares.medium.com/membership
– Follow me on Medium: https://lucas-soares.medium.com/
– Twitter: https://twitter.com/LucasEnkrateia
– LinkedIn: https://www.linkedin.com/in/lucas-soares-969044167/
– Tiktok: https://www.tiktok.com/@enkrateialucca?lang=en
If you liked this video don’t forget to like and subscribe! Thanks and see you next time! 🙂
Thanks for this helpful resource. In the showPaperSummary function what is the "organisation key" for openai.organization ? Where can one find it? Thanks again
great video!!! can you share the code? as it is quite different from the one in Medium (which for me does not run..)
Thanks for the video! FYI that your audio is wonky and hard to listen to (especially faster than 1x speed).
Hey, thanks for the video! I get the error: "No such organization: API KEY org."
How can we know the API key organization?
Is it possible to summarize a big article, but I would like to keep specific words??
In their website, they seem to clue the algorithm to spit out a summary rather than a continuation, am I correct?I did in Node JS but was curious about others approach
Thank you very much!
I get the following error:
paperContent = pdfplumber.open(paperFilePath).pages
NameError: name 'paperFilePath' is not defined
any ideas?
How does this deal with the 4000 character limit of GPT-3?
Thanks so much for this. It helped me so much and I learned so much. I took a little bit of time and I worked on my own script so that I could make sure that complete sentences were being sent to open ai to obtain the summary. I also found that by adding a prompt asking open ai to summarize the text that I got much better summaries. I hope you find it useful. Oh I used pdfplumber separately and just outputted one big txt file with all of the text so the code looks a bit different because I used pdfplumber separately.
#importing the necessary modules
import re
import os
import openai
def showPaperSummary(text):
tldr_tag = "n tl;dr:"
openai.organization = 'org-eSJ3QB0DS0hsSii3ov4jDGI5'
openai.api_key = "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"
#engine_list = openai.Engine.list()
prompt = "Please summarize this text"
# for page in paperContent:
text = text + tldr_tag
print("