openAI/ChatGPT: Using embeddings for training your dataset based on organizational documents (FULL)
A very common use case from chatGPT is to know how to perform questions based on existing documents. It can be very important when the chatGPT answers needs to be related to a specific context, and not in the general internet knowledge.
For this purpose this video shows a step by step process to achieve this result using python programming language.
Hope you enjoy the video and can utilize this content for helping you in the achievement of you personal and professional goals.
Knowledge is one of few things we can share without reducing or subtracting it. This channel is an attempt to bring education about technology, without costs, to all corners of the planet. We believe technology knowledge in good hands, like yours, can assist to build a better world.
Do you want to support this channel? Please just subscribe.
Link on github:
https://github.com/rafaelqg/chatgpt_c...
00:00 Embeddings and its appliction context training fles
0:59 Welcome and video overview - the 3 modules/parts of the presented solution
2:02 Python code presentation consuming openai api for chatpt
3:33 Genereate embeddings from a file
4:06 Creating an array with embeddings for all trained files.
6:53 Receive user questions and find the biggest similarity (cosine) with trainned documents (embeddings)
8:29 Send question to chatgpt with similar documents content as context of a conversation
9:40 Running the application and evaluating results
English subtitles available!!