Scribd
Upload
6 views

Section 060 Tips Tricks and Resources

Uploaded by

nanujune2004
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
6 views

Section 060 Tips Tricks and Resources

Uploaded by

nanujune2004
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 33

Follow us on https://kodekloud.

com/ to learn more about us


Hello and welcome back. So before we take a deeper dive into Lang Chain and some of its core concepts, I want to give you
some tips and tricks plus some resources that will help you navigate this complex maze of the Lang Chain ecosystem. So
let's first look at understanding the libraries.
So Langchain has multiple libraries. It is one of the fastest growing ecosystem and obviously there are dozens of libraries
that are available within the langchain family and the langchain ecosystem. So as a developer, how do you look at the land
chain library hierarchy? Well, you need to understand three different libraries and these are essentially what we are going to
deal with going forward. The first one is langchain core. So think of land chain core as the runtime. So Lang chain is evolving
almost like a platform. So every platform has a runtime. So Lang chain core comes with the runtime that is responsible for
executing the chains, agents, tools and pretty much everything that you run at the at the high level at the top, top, most
layer of the stack. It also has base abstractions and interfaces. So think of it like the actual definitions that are going to be
implemented by community and by Lang chain itself. So it has some base classes and abstractions and then it
has the most powerful piece of Langchain called LCEL Lang chain expression language. And I am going to walk
you through this in the next lecture. But to summarize, Lang chain core contains the runtime, the base classes,
base abstractions and L cell Lang chain community essentially inherits what is available in langchain code and
builds upon it.

(This is the script for slide 3 to 8)


For example, Open AI has a library exclusively to integrate with langchain which is langchain Open AI, and similarly other
LLM providers like Anthropic Cohere and Amazon Bedrock, Azure Open AI, Google Virtex AI.So there are dozens of open
source and commercial LLM providers.They when they integrate with land chain, they come back with their own library.And
that library is a concrete implementation of the base abstraction defined in land chain core.So land chain community is not
just confined to LLMS.There are vector databases, there are retrievers, there are loaders, there are tools.There is a huge
ecosystem around the Lang chain community and all of them fall under this third party integration Bucket.And then Lang
chain is the library that is implementing some of the base classes defined in Lang chain core and implements some of the
most essential building blocks that you do not need to essentially depend on 3rd party.So chains, agents,
retrieval strategies that are very very important and that also become the part of the applications cognitive
architecture which is the brain.So this has the core logic and the concrete implementation of the abstract
classes and abstract definitions available in Lang chain core.So the best way of visualizing this is Lang CHAIN
CORE is like the runtime plus base classes and it branches out into Lang CHAIN community left to the
ecosystem to build their own libraries and then Lang chain which will build upon this to give you the actual
implementation of the base abstractions defined in Lang chain core.So this will become clear when we take a
look at the Jupiter notebook in the next demo.

(This is the script for slide 9 to 15)


So having understood the core libraries, I want to introduce some tips and tricks that will help you debug langchain
applications and also get an insight into what is going on when you are executing langchain based applications. Langchain
can be overwhelming for beginners and it might also look like a black box because you really don't know what is going on
behind the scenes. You will create a chain from prompt to LM to an output parser, but you have no visibility into what is
going on South. If something really goes wrong in terms of the formatting or in terms of the responses, or you want to just
understand the entire execution life cycle, you need to know some tools. You need to understand some tips and tricks to
really gain insight into the runtime execution. So in this lecture I am going to walk you through some techniques that will
help you debug your applications and also get visibility into what is going on South. In the next demo I will
introduce you to the debug, the verbose flag and implementing a very powerful concept called callbacks. So
let's switch to the demo.

(This is the script for slide 17 to 19)


So having understood the core libraries, I want to introduce some tips and tricks that will help you debug langchain
applications and also get an insight into what is going on when you are executing langchain based applications. Langchain
can be overwhelming for beginners and it might also look like a black box because you really don't know what is going on
behind the scenes. You will create a chain from prompt to LM to an output parser, but you have no visibility into what is
going on South. If something really goes wrong in terms of the formatting or in terms of the responses, or you want to just
understand the entire execution life cycle, you need to know some tools. You need to understand some tips and tricks to
really gain insight into the runtime execution. So in this lecture I am going to walk you through some techniques that will
help you debug your applications and also get visibility into what is going on South. In the next demo I will
introduce you to the debug, the verbose flag and implementing a very powerful concept called callbacks. So
let's switch to the demo.

(This is the script for slide 17 to 19)


So before going ahead and proceeding to the next set of lectures and lessons in this course, I want to put a disclaimer.
Langchain is a rapidly evolving framework, platform or library, you can call it any of this. With the pace at which the
generative AI is evolving, Lang Chain is equally growing at the same pace. So this course is specifically based on Langchain
0.1.1.XI put an X here because when I started recording this it was 10 and it is already 11. So to make sure that all the
Jupyter notebooks and the code work the way they are, make sure you are basically using the Lang chain library that
matches this version which is oh point 1.11 or 0.1.10. So with that in place, I also want to talk to you about some resources
that will help you stay current with the pace at which Lang chain is evolving. So there is Lang chain website and then there
are docs. You can also access the Python SDK and then subscribing to Lang chain blogs will actually help you
with that.
Hello and welcome back. So before we take a deeper dive into Lang Chain and some of its core concepts, I want to give you
some tips and tricks plus some resources that will help you navigate this complex maze of the Lang Chain ecosystem. So
let's first look at understanding the libraries.
LangChain is evolving rapidly. This course is based on 0.1.1x. Refer to the resources to stay up to date:
LangChain Website
LangChain Docs
Python SDK Docs
LangChain Blogs
LangChain is evolving rapidly. This course is based on 0.1.1x. Refer to the resources to stay up to date:
LangChain Website
LangChain Docs
Python SDK Docs
LangChain Blogs
Follow us on https://kodekloud.com/ to learn more about us

You might also like