Building Intent Based Advanced Semantic Search From Scratch
In this video, we are going to build advanced semantic search from scratch by using new updated elastic search and the new json mode of chat GPT. If a user searches for: Blue T shirts for men under 2k. Normal semantic search results might give results which contain other colour T-shirts as well. It can contain T-shirts for women as well and it will most probably not understand that we want T-shirts only under the price of 2000. We will see how we can fine-tune these results so that the user will get most accurate products according to his/her query
Tech Stack:
New json mode of GPT 3.5 Turbo Model
Elastic-search 8.11 ( which supports vector dimensions up to 4096)
OpenAI’s text-embedding-ada-002 model
Github of entire project: https://github.com/abidsaudagar/Advan...
Chapters
00:00 Introduction
00:21 What's new in Advanced Semantic Search
01:33 Architecture of entire project
03:05 Let's start coding
03:16 Dependencies and Data Cleaning
03:48 Converting text to vector using openAI's embedding
05:14 Inserting data to ElasticSearch
07:42 Searching data from Elastic Search Index
09:15 Applying filters on semantic search results
10:41 Getting intent of users using GPT 3.5 turbo model
16:06 Testing results of advanced semantic search
16:37 Important update & Thank You
Other references mentioned in the video:
https://platform.openai.com/docs/guid...
https://platform.openai.com/docs/guid...
https://openai.com/pricing
Link of my previous semantic search video:
Build Semantic-Search with Elastic search and BERT vector embeddings: • Build Semantic-Search with Elastic search ...
Checkout my similar videos:
1. Coding a privateGPT using LANGCHAIN, HuggingFace Embeddings and FREE LLM: • Coding a privateGPT using LANGCHAIN, Huggi...
2. Coding chatbot to talk with QURAN using ChatGPT and Langchain: • Langchain + ChatGPT - Coding chatbot to ta...
YOU CAN FIND ME HERE AS WELL :
My Instagram: / abidsaudagar
My LinkedIn: / abidsaudagar
/ building-advanced-semantic-search-with-ela...