Brandon, I am nearly 70 years old and I have been a programmer for over 4 decades. I started as a Cobol programmer and been updating my skills as I went along. Never has been my learning at such a rapid pace as it is in the last year or so since all the AI tools have been arriving. I had just a basic knowledge of Python earlier, but now I have been able to do wonderful things with what I can develop with AI agents using tools available. Your tutorials have been the best I have seen so far. I can't thank you enough for providing the complete source code and explaining very well everything in your videos, including the environment setup. Keep up the great work!
@padhuLP WOW thats awesome that you are still doing and learning at 70. You must be an amazing person, more young people meed to take some tips from you on life lessons. Even better I assume you are retired and doing all this for fun? I only hope I can live and learn as long. Keep be inspirational and take care for your future, Love from N.Ireland ❤
Wow, this LangChain Crash Course was amazing! It broke down the concepts way better than the official documentation. I learned so much from this video, I just had to come online to say thanks! 😂 Really clear and to the point. Please keep making more videos like this on LangChain, LangGraph, and similar stuff! Thanks again! 🙏🙏🙏
Surely one the best tutorial videos I have come across on youtube so far. Not just for Langchain.....but across all the video tutorials for any technology I have seen so far......This is precise, crisp... hits the nail on the head with the learnings!!! and the best thing is it packs so much in.. like a crash course.. not wasting any time at all on typing the code during the recording... saves a lot of time and does not digress the listeners attention!! @brandon you are a good teacher and you know how to hold attention of students like me who sometimes never go beyond 10 mins into the video if the teacher is not able to hold attention and present him/her self well !!! SUPER JOB!!! FORMAT of the Training EXCELLENT!!!
thanks a lot for you help bro . This video was really useful i had got stuck the whole time with outdated langchain . I watched many videos and learn't only outdated half stuff but all thanks to you i have learned langchain completely
Thanks man! It’s funny because I think all of us are getting stuck in their docs! This tutorial was born out of frustration 😂 Once you get everything working, LangChain is pretty great!
Great video! I just can't comprehend how others can have such a long video and not explain concepts and ideas in the way that you do in 5 minutes. xD ty!
you deserve more subscribers Brandon. "A" class content. Can u do a tutorial on fine-tuning pre trained models and also about GraphRAG? Thank you in advance
Brandon, your tutorial is the best one I've come through in RU-vid about langchain. Thorough and easy to understand and follow! Both concepts and coding logic well explained! You deserve more subscriptions!
I can’t thank you enough for the time and effort you put into creating such high-quality content. It’s clear that you’re passionate about helping others, and your hard work doesn’t go unnoticed. Looking forward to more amazing videos from you! 😊 chain = great_video | detailed_explanation | grateful_for_effort | look_forward
I think OpenAI still gives out some free credits when you sign up. But, eventually you will have to pay. It's only a few pennies per request so it's not terrible.
I like what people are doing with AI, this guy built a text to image app in this video: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-4LnycOUhqyc.html Would you be able to cover this? And also has a lot of langgraph content, with a different perspective.
Great suggestion! I’ll definitely add a LangGraph master class to the queue for you guys! 😁 And I created this tutorial because the docs for the new version of LangChain are kinda confusing. I’m hoping this video will help everyone spin up on the new version! If you have any questions on anything, please let me know!
@@bhancock_aiYou're right! I started working with LangChain about a year ago, and now that I need to use it for my current project, almost everything has been deprecated. It was very frustrating at first. Also your CrewAI+Groq video is very good for learning purposes.❤
@@bhancock_ai Yes plz and show how to use function tool calling and react agent with chathistory in it able to use doc store to retrieve information and plz with open source model like chat groq that after making the application can be swapped with gpt so on that structure
Why all projects use OpenAI, they removed the free credits and students don't have money to invest to learn. So kindly make use of some open source model also.
That is a bit lazy. You can easily find out how to change the model and embedding provider in the official documentation and everything else applies. He explained how to switch to Anthropic, and its totally analogous. PS: Im not even a programmer. I am a self-taught and my background is in medicine, so I don't think I have any comparative advantage.
I think it just too easy to use open air and it works.. ideally the youtube community ahould make an effort to give altenrative options so they can build something cheaper.. for end users. I cant imagine building and app with open Ai inference it would be so expensive for my customers.. it just bad proposition..
I've been creating new keys but still getting this You exceeded your current quota, please check your plan and billing details., anyone knows how to solve this?
Brandon, this LangChain tutorial is hands down the best I've seen! You nailed all the fundamentals and went deep into the details in such a human and relatable way. Your style makes complex stuff super easy to understand, and the practical examples are spot on. Love that you didn't push any products, just pure high-quality content focused on real value!! This is a must-watch for anyone diving into the matter. Huge thanks for creating such an awesome resource!!! I'll never look back at any other framework after understanding the power of LangChain. I hope you'll create a tutorial with the same level of detail for the new LangChain GenUI. Your style makes learning so much more enjoyable and effective! Thanks for all the great content!
Brandon, this course is outstanding! It's exceptionally well-structured, clearly explained, and superbly produced. I aspire to create tutorials to my channel with half of the quality you've got here :)
this is a great tutorial. Thank you! Can you please advise what tool do you use to draw your diagrams in real time in the video? for example at minute 49:00 you demo chaining concepts
Brilliant tutorial which explores to the depths! Btw Brandon, how to incorporate google gemini or groq or ollama instead of openai service? The prompt template is different for each case. Please make a tutorial on incorporating the same!
I have been using Ollama while working through this tutorial with very few issues and I just used the prompts as they were in the code base. Give Ollama a try. I think you will be surprised at how easy it will be to swap in llama3.
Super cool Video Brandon.. excellent content and nice pacing. Thank you so much 🙏 Just completed the Chat Models section. excited to continue with the rest of the sections. For me Gemini model API works and they are quite generous with their Free Tier plan compared to OpenAI
Great tutorial... Now I started to understand better Langchain... I bought one Langchain course from Udemy and after watching the first section I just got lost and frustrated... Your explanation is far better... Thanks a lot... Liked and subscribed... definitely I will watch more videos from you.... Thanks again!
Trying to follow your instructions for firebase chat history saving, did all exaclty the same but the chat-history collection (there is no word in a video if I should create one myself or it would be created automatically). Starting it with python command, and get TypeError: cannot pickle 'classmethod' object error message. What could be the issue?
would like to see a version that uses only open-source components, that can be run locally - at the very least provide some pointers to these alternatives
This is a great class! One question. I've been following using Azure OpenAI Studio deployments and have been just fine until the RAG portion. Even at 320k TPM on my embedding I still get rate limited, and consequently timed out for 24hours, as the amount of tokens required for RAG 2a ends up being significantly higher. Do you know of a way to properly programmatically set a rate limit so I don't get a paddling from Microsoft? Thanks!
Hi Brandon, thank you for this great tutorial. I have gone through each word and hands on. Looks like Tool Decorator explanation is missing. Could you please check once?
Hi thanks appreciate the tutorial! Can I give you some feedback? I cannot stand the sound effects you've included -- the swoops and pops are 🎉 distracting and unnecessary. Maybe it's just me, and I can just watch it on mute 😅💥
Hi Brandon, going through your tutorial I saw that once you copy paste the location from poetry shell the squiggly lines should disappear but for some reason this isn't working for me as it says the dotenv module is not found. Thanks for providing this resource!
Poetry. do you have a plane requiments.txt file. I am not a fan of poetry it cause issues with my machine by not releasing the lock and freezing my machine.. I find just using a reequipments.txt for adding my python pip loas to be simple and does not add additional overhead and running another tool, poetry, to manage .
I am not clear at 1:01:47 How did you get to print the output of the uppercase_output? It was further chained to count_words. The output that got printed should be just word count. I am confused.
Hi Brandon, Thank you so much for putting together this Master Class. Previously, I was watching videos of older Langchain versions and encountered many changes and warnings when coding. Those videos had numerous gaps. After watching your videos and coding along, I’ve found it much easier to follow the entire Master Class. Your explanations are very clear, and your visual diagrams greatly help in understanding the flow. Thanks again!
Excellent couurse. I join Langraph's requests. Maybe Langflow would also be nice (I think it's very practical) why not include the groq API? I think it's free. I have not be able to install it throught Poetry in the directory. Congratulations on the course.
Brandon, in example 7_rag_conversational.py, for the {context} in the qa_system_prompt template, is it something coming from the previous Runnable from the chain? I don't see it being as one of the input parameters.
I stumbled over the same question. Tweaking things a little bit I found the answer: The create_stuff_documents_chain(...) expects as input a dictionary that must have a “context” key that maps to a List[Document], and any other input variables expected in the prompt. [From the langchain V0.2 documentation on create_stuff_documents_chain(...)] It seems that there is an optional argument (..., document_variable_name: str = 'context') to change the "variable name to use for the formatted documents in the prompt. Defaults to “context”."
after 46 mins into videos. amazing man ❤️🔥. i really needed this. this is the best resource to learn langchain i came across. full review after finishing the video!!
Hey! I’d love to do some deeper dives into LangChain and agents! What else would you like to learn? If you’re interested in using a multi-agent framework, I definitely recommend trying out CrewAI!
Hey! Poetry is a tool that allows you to easily manage your python environments so you shouldn’t have any issues using it with Pycharm. Also, how do you like Pycharm compared to VSCode?
@@bhancock_ai - Thanks Brandon. I like PyCharm as it comes with many built-in tools and features, such as code analysis, debugging, and testing. But community version has some limitations like we cannot run ipynb. And it can be heavy on system resources, which might slow down performance. But running mini RAGs is not an issue. I also switch to Jupyter for ipynb .Once again thank you for your great compilation on langchain. Looking forward to trying out all projects.
Hey! If you add your email, it will send you an email that contains links to the source code plus links to source code for the rest of my projects! Please let me know if it’s causing you any issues!
At 59:00 in under the hood, this way of writing feels more intuitive to me: `invoke_format = RunnableLambda(lambda x: prompt_template.invoke(x)) invoke_model = RunnableLambda(lambda x: model.invoke(x)) parser_output = RunnableLambda(lambda x: StrOutputParser().invoke(x)) chain = RunnableSequence(invoke_format, invoke_model, parser_output)`