Скачать или смотреть Building 🦜🔗Langchain RAG App from Scratch | 🎬 Spoiler Blocker(Part-1)

Labels on IMBD are not ground truth as-
1. Data was scraped in 2018-19 from imbd website. Not sure about what model they were using in pre-transformer ERA to label as spoilers.
2. Assuming imbd model to be perfect(less probablity), since it uses RL logs, it gonna be very unstable as spoilers depend on individual to individual.
3. Training model using artificial labels(non ground truth) will be difficult to converge.
4. So using prompt engg, RAG & LLM to devise the OG way to solve this in unsupervised fashion.

Summary- spoilers are the text reviews which reveals major twist and turns in movies and kills viewers experience. Using capabilities of Generative AI lets build a simple RAG langchain app to solve spoiler blocking. This app will scrap the data in real time whenever user clicks imbd website and add the movie summary & synopsis into vector db to perform retrieval augmented generation(RAG).


Code- https://github.com/Guggu-Gill/spoiler...

Dataset link-https://www.kaggle.com/datasets/rmisr...
Chapters-
1️⃣ Introduction to Problem & Demo run - 0:00

2️⃣ Basics of RAG & Vector DB - 2:37

3️⃣ Architecture of RAG App - 6:39

4️⃣ Skelton Code walkthrough - 9:17

4️⃣ UI Skelton- 17:50

Next video releasing soon with automated chrome extension & deployment of chains.

Contact-
  / basantsingh1000  
  / basant_singh_22g  

Buy me coffee 😁- https://www.buymeacoffee.com/singhbasant