This video is part of my Best Selling Udemy Course on Amazon Bedrock and Generative AI - www.udemy.com/course/amazon-bedrock-aws-generative-ai-beginner-to-advanced/?referralCode=A02153572B8864F928E7&couponCode=NVDPRODIN35
@chloe-z5f
Ай бұрын
Hi, great video! do you have the code uploaded as a repository on github?
@alpha_ray_burst
4 ай бұрын
Thanks for this video. It's exactly what I needed! I'm curious though... what is the purpose of using Anaconda to launch VSCode? Is that different from launching VSCode by itself?
@trisalrahul
4 ай бұрын
Thank you… sometimes streamlit gives issues if not launched through anaconda…
@khandetanvikram
4 ай бұрын
Hi, It was a good session, can we get the git link pls
@trisalrahul
4 ай бұрын
Thanks Vikram...This video series is part of my Udemy course on Bedrock which you can check from the description...
@muhammadmubashir1123
5 ай бұрын
Great knowledge. Can I get the code and other sources shared in this video ??
@trisalrahul
5 ай бұрын
Thanks…will try n upload somewhere…you can also try my Udemy Course if you may like..course link in the description…
@sudiptapalchowdhury2828
4 ай бұрын
Hi Rahul, could you please guide me as to how to fine tune such application using RAG, Bedrock, FAISS. For example loading and indexing large size documents into vector takes much time, how this can be reduced...secondly, querying and getting a response also takes time, what can be done here to increase performance of query. Will appreciate if you can provide your valuable suggestions
@trisalrahul
4 ай бұрын
Hi Sudipta - It a good question but too complicated to answer in chat...Will depend on various factors like what is the use case, complexity etc...We know for sure that there is latency in LLM's due to large number of parameters, then what what kind of Vector DB you use how to index etc...these days one approach is to use Small Language Model which have lower latency due to lesser number of parameters...but overall you need to look at all these considerations and design the solution and experiment through a PoC to get optimal solution...
@continuouslearner
6 ай бұрын
Great video. At the appropriate time in the video, can you please link (as a popup youtube overlay link which we can tap) to your other video about vectors and vector db? Also, what is the difference between using sagemaker studio jumpstart and bedrock?
@trisalrahul
6 ай бұрын
Thank you @continuouslearner ...Sagemaker Jumpstart - you need to provision the undelying infrastructure for the LLM's while as in Bedrock - its a serverless service - no infra provisioning or management and pay as you use...
@trisalrahul
5 ай бұрын
Here is the link to basics on vectors and vector db search kzitem.info/news/bejne/24ZrrGqDhp6VmYY
@sudiptapalchowdhury2828
4 ай бұрын
How can I expose an endpoint for this back end, so that an external application can send input query and the backend returns generated response. Thanks for your session, it was great to learn!
@trisalrahul
4 ай бұрын
Thanks Sudipta for watching - There are many ways to do it...The easiest and serverless way to do it is use API GW, Lambda and Bedrock with KnowledgeBase in AWS...Take a look at other video part of the playlist where i demonstrate this use case...
@sudiptapalchowdhury2828
4 ай бұрын
@@trisalrahul thanks for the reply Rahul, I meant will I be able to expose the same backend written using Langchain from Lambda using Python?
@trisalrahul
4 ай бұрын
@sudiptapalchowdhury2828 - Yes, you will be able to do that….langchain works with lambda …though you may additionally want to include api gateway to secure your endpoint - Maybe in future i will create one use case on lambda + langchain …
@chloe-z5f
Ай бұрын
Hi, great video! do you have the code uploaded as a repository on github?
Пікірлер: 19