PR for merging content on Notebook for RAG with Llama 3 open-source and Elastic #260

[rishikeshradhakrishnan]

Notebook for RAG with Llama 3 open-source and Elastic #260

[gitnotebooks]

Found 2 changed notebooks. Review the changes at https://gitnotebooks.com/elastic/elasticsearch-labs/pull/266

[joemcelroy]

Commented on notebook notebooks/integrations/llama3/rag-elastic-llama3.ipynb Cell 4 Line 1

!pip install llama-index

nitpick: update this to be one line for pip install, similar to how the other notebooks have done this

Respond and view the context here.

[rishikeshradhakrishnan]

Updated the code.

[joemcelroy]

Commented on notebook notebooks/integrations/llama3/rag-elastic-llama3.ipynb Cell 6 Line 2

from llama_index.core.node_parser import SentenceSplitter
from llama_index.core.ingestion import IngestionPipeline

nitpick: move the imports to where the code being used, so its obvious which cell needs which import.

Respond and view the context here.

[rishikeshradhakrishnan]

Updated it accordingly.

[joemcelroy]

Commented on notebook notebooks/integrations/llama3/rag-elastic-llama3.ipynb Cell 13 Line 1

## Execute pipeline, 

remove ,

Respond and view the context here.

[rishikeshradhakrishnan]

Updated this as well.

[joemcelroy]

Commented on notebook notebooks/integrations/llama3/rag-elastic-llama3.ipynb Cell 15 Line 1

The embeddings are stored in a dense vector field of dimension `4096`. The dimension size comes from the size of the embeddings generated from `Llama3`.

is that right? the embedding is 4096 dims? Could we improve this with picking a smaller dim model instead if so.

Respond and view the context here.

[rishikeshradhakrishnan]

AFAIK and checked, there didnt seem to be any option to change that.

[joemcelroy]

Commented on notebook notebooks/integrations/llama3/rag-elastic-llama3.ipynb Cell 1 Line 1

# RAG with Elastic and Llama3

this examples are either LangChain or Llama-index. I would update it like so and be obvious that its using these frameworks behind the scenes.

Respond and view the context here.

[rishikeshradhakrishnan]

Updated it.