Can you please explain the prompt flow with multivector-multimodal RAG which includes embedding , indexing and inference of text chunk, table chunk and an image chunk.
thanks a lot for this wonderful video. every step is very clearly mentioned. only thing missing is the function "generate embeddings", i checked your medium article, it is also missing they function. can you pl. update the medium article or upload the notebook to the github pl?
I have a question. like for data, you take it from your local, my use case is I need to do a api call to fetch the data. I am not clear on the part how to achieve it. do I need to write a separate code which will hit another api, get data from the response, embed it and then save it in azure ai search? Can you plz guide me on this. Rest all is same as your use case, it is just the data we are getting from third party api.
If possible can you do embedding to store PostgreSQL flow, or best way for storing embedding…. Or is it good to store temporarily on Faiss and perform similar search 👀
Excellent Explanation and step by step guidance 🤗
Glad it was helpful!
Can you please explain the prompt flow with multivector-multimodal RAG which includes embedding ,
indexing and inference of text chunk, table chunk and an image chunk.
Yup, it’s in radar. Stay tuned!
Please make video on AI studio Indexing which is used in Azure ML prompt flow
Will do soon 👍
thanks a lot for this wonderful video. every step is very clearly mentioned. only thing missing is the function "generate embeddings", i checked your medium article, it is also missing they function. can you pl. update the medium article or upload the notebook to the github pl?
I have a question. like for data, you take it from your local, my use case is I need to do a api call to fetch the data. I am not clear on the part how to achieve it. do I need to write a separate code which will hit another api, get data from the response, embed it and then save it in azure ai search? Can you plz guide me on this. Rest all is same as your use case, it is just the data we are getting from third party api.
Yes, you’re correct. Call your third party API, grab the response and then you can push that to Azure AI
Can we Use existing index from AI search instead of creating new?
Of course, you can.
If possible can you do embedding to store PostgreSQL flow, or best way for storing embedding…. Or is it good to store temporarily on Faiss and perform similar search 👀
Hey, thanks for the suggestion. I have added this to my list. Will do it soon.
@@shweta-lodha awesome, Thanks
On it
Stay tuned