Several days passed since I am trying to get out of this infinite loop of tutorial hell This channel is one of the best Myself also got overwhelmed by the information and this man teaches things so well and felt like understanding every concept he teaches .. so it was so much information for me and I saved the video and stopped it for now …❤
Thanks, excellent walk through !! Have a quick question. From the documentation, seems it currently only supports openai and anthropic. Can you pl share some pointers for huggingface models support. Thanks again.
Thank you for your tremendous work! Maybe you could give me a hint how to achieve such process: 1. User message as an input. 2. Message type classification [General, Action request]. 2.1. If General -> forward the query to a React agent with basic tools and chat memory. 2.2. If Action request -> summarize the request with a parsed output chain -> forward the output to another chain along with a retrieved context for execution. I can't use a simple react agent with plenty of tools for that, because without the classification and summarization step, agent fails to understand what to do. I must distinguish some basic stuff from the complex, dedicated stuff which may involve many steps to execute properly. Ofc, I could have two or more agents and determine which to ask in the beginning, but it's not what I want to achieve. Thank you!
Do you think you can create a prompt classifier? Your classifier should understand that class of the query/prompts. Once you classify the prompt, according to class... You can call the respective functions. Does it make sense?
@@AIAnytime Thanks for the insight. Yes, I was thinking about using the classifier in the first step, although it doesn't walk around the problem of agent executor. It looks that it needs to be declared prior to the classification step. So, this way all of the chains will execute as a part of the react agent chain and thus I'm not sure if it will work. Did you try something similar, or maybe you could consider making a video about more complex LCEL designs? Thank you!
Thanks for sharing the video. But a factual correction - OpenAIEmbeddings is a class , OpenAIEmbeddings() is constructor to create object. It's NOT function. Similarly in other places
![LangChain Master Class For Beginners 2024 [+20 Examples, LangChain V0.2]](http://i.ytimg.com/vi/yF9kGESAi3M/mqdefault.jpg)
Several days passed since I am trying to get out of this infinite loop of tutorial hell
This channel is one of the best
Myself also got overwhelmed by the information and this man teaches things so well and felt like understanding every concept he teaches .. so it was so much information for me and I saved the video and stopped it for now …❤
Man thanks again for a great video, now I’m getting lost with so much knowledge. Thanks again for valuable content.
Glad to hear it!
Most underrated channel. Hats off! keep up the good work.
Thank you so much 😀
Thanks, excellent walk through !! Have a quick question. From the documentation, seems it currently only supports openai and anthropic. Can you pl share some pointers for huggingface models support. Thanks again.
Tried using Llama 2, works perfectly.
Hi, Sorry, I'm a little new in this stuff
Can I use the output of one chain as the input of second chain, in a bigger chain essentially?
Thank you :)
Yes, a chain is also a LCEL object so mutiple chains can be combined in the same way.
Thank you! 🤓
Thank you for your tremendous work! Maybe you could give me a hint how to achieve such process:
1. User message as an input.
2. Message type classification [General, Action request].
2.1. If General -> forward the query to a React agent with basic tools and chat memory.
2.2. If Action request -> summarize the request with a parsed output chain -> forward the output to another chain along with a retrieved context for execution.
I can't use a simple react agent with plenty of tools for that, because without the classification and summarization step, agent fails to understand what to do. I must distinguish some basic stuff from the complex, dedicated stuff which may involve many steps to execute properly. Ofc, I could have two or more agents and determine which to ask in the beginning, but it's not what I want to achieve.
Thank you!
Do you think you can create a prompt classifier? Your classifier should understand that class of the query/prompts. Once you classify the prompt, according to class... You can call the respective functions. Does it make sense?
@@AIAnytime
Thanks for the insight. Yes, I was thinking about using the classifier in the first step, although it doesn't walk around the problem of agent executor. It looks that it needs to be declared prior to the classification step. So, this way all of the chains will execute as a part of the react agent chain and thus I'm not sure if it will work.
Did you try something similar, or maybe you could consider making a video about more complex LCEL designs? Thank you!
Great idea, let me plan. Thanks
Trying to follow along but keep on getting error messages...
You need to update Langchain. And then restart the kernel of the notebook.
Thank you! Will give that a try! @@AIAnytime
Thanks for sharing the video. But a factual correction - OpenAIEmbeddings is a class , OpenAIEmbeddings() is constructor to create object. It's NOT function. Similarly in other places
Thank you!
Welcome.
Thanks bro
Welcome
watching someone type out what should be a prepared slide is a waste of time