What is LangChain?
2024 ж. 14 Нау.
110 555 Рет қаралды
Learn about IBM watsonx→ ibm.biz/BdvkK8
LangChain became immensely popular when it was launched in 2022, but how can it impact your development and application of AI models, Large Language Models (LLM) in particular. In this video Martin Keen shares an overview of the features and uses of LangChain.
Get started for free on IBM Cloud → ibm.biz/sign-up-now
Subscribe to see more videos like this in the future → ibm.biz/subscribe-now
After watching a ton of courses on coursera developed by prestigious universities, and tech companies like google, meta, etc., I found IBM' courses are the best: comprehensive, clear and concise! Thank you
Just want to thank the wonderful presenters for this excellent series and IBM for its commitment, innovation, and broad contribution to education in this field.
What innovation you're talking? The community innovation?
I appreciate presenters and mentors who make concepts easy to understand, as I am not a native English speaker.
Best description of LangChain I have watched
Pure gold! Thanks a lot for this. Best regards from Nairobi, Kenya
hey i want you to contact me
Great topic and great presentation of it -- a good breakdown of what it is and the use-cases it tries to fulfill. I like how you filled in some of the text instead of writing it out -- saves a lot of time. Thanks.
great work for this complicated frame work!
martin KEEN(nominative determinism at work?) is my all-time fav IBM presenter! THis one no exception Great, clear expo of LangChain..
As soon as I saw him start writing backwards, I lost all concentration on the presentation and was wondering "how did he do that", and was figuring it out instead of listening to the content.
A complex topic explained clearly.
Great video! I just published a chat bot for home brewers, but it’s a bit basic so a friend of mine recommended using langchain to have it provide better responses and take meaningful actions. This was really helpful for me to get started, thank you!
Ooooh share a link if you can!
Yeah could you share
This was super useful and easy to understand, thank you so much for posting!
Great explaination. Thank you
Very well explained ! Thx
Thank you, Great tutorial, and easy to understand manner. We would like to ask you to explain about Lang chain concepts like indexes and chains. thank u again.
Thanks for sharing your knowladge with us. Thanks from Brazil.
Way to break down the chain. Well done.
As always informative and entertaining at the same time. Would be great if you could show us some demo applications as well as some high level sample integration code or component. Thanks.
Best explanation.
very helpful, thanks!
Thank you very much for this introduction! I found it really useful and helps me navigate all the information and new programs around AI. Would you be able to provide few examples on what can we built on langchain ?
thank you
Can we expand the capacity of input prompt using langchain if yes can you suggest any methods??
In the same program can we query one LLM (Like Llama2) , get response & use that response in a chain and pass that to another LLM (like google gemini)... All the in same program?
LangChain exists a long time and its community is huge. Nice to finally get an intro from IBM about that long existing technology and community product.
Can you use this tools to create a chatbot for internal databases and taking care of the data privacy? For example, I have my documents with private information and I want to use Langchain to ask question about it... Can I create it with langchain without exposing my data?
How does langchain stack against llama index? Which is most preferred?
Best explanation. Please can IBM provide end to end AI courses that can make us market ready for AI job or creating AI start up. So many noises out there
Most people use langchain to extend memory on LLMs, do you have an alternative for that?
Ibm explaining langchain. I guess langchain is getting production ready. During the start langchain was a bit rough.
- [00:00] 🛠 LangChain Overview - LangChain is an open-source framework for developing applications utilizing large language models (LLMs). - Components of LangChain include abstractions, LLM modules, prompts, and chains, streamlining the programming of LLM applications. - [01:26] 🔄 Abstractions in LangChain - Abstractions in LangChain simplify complex NLP tasks by representing common steps and concepts necessary to work with language models. - These abstractions can be chained together to create applications, minimizing the amount of code required. - [02:22] 🔗 Components of LangChain - LangChain comprises LLM modules, prompts, and chains, enabling developers to integrate different models and execute sequential functions to achieve desired tasks. - Components like prompts formalize the composition of instructions given to LLMs, facilitating the development process. - [03:47] 📊 Data Handling in LangChain - LangChain supports various data handling mechanisms, including document loaders and text splitters, enabling access to external data sources and efficient text processing. - Vector databases and memory utilities enhance data retrieval and management within LangChain applications. - [06:09] 🤖 LangChain Use Cases - LangChain facilitates diverse applications such as chatbots, summarization, question answering, data augmentation, and virtual agents. - Integration with existing workflows and robotic process automation (RPA) enhances the functionality of LangChain.
cool video)
Just to clarify, LangChain ~= JSON but for LLMs?
Does he write inversely on the board?
👏👏
Um.. why is brulosophy guy on IBM's channel explaining langchain??
LanguageModelChaingang ;)
Am I the only one who wonders how this presentation actually works? Is he writing on a mirror? If he wrote on glass while standing behind it, it would be in reverse for us... How does some of the words autocomplete? How does it glow? Is it magic?
It’s actually a thing they invented to do this, I wondered the same thing and the actual ibm channel responded with a link to how they do it. Pretty sure it’s the video flipped horizontally. While he writes on a glass ( I cant remember anymore it was a long time ago the ibm channel responded)
I think he writes on a glass, and the camera is behind the glass. When the video is shot, it's mirrored so that left becomes right, and the right becomes left. Glowing is not a problem, there are fluorescent paints.
The setup's left side view will be like: Person |Glass| Camera Glass will be big enough to cover the frame as in the video. Face is illuminated by lights fixed on the glass's frame. This will avoid reflections which would have been seen on the glass if illuminated from the front side. He writes on the glass with chalk pen. Video will be recorded on to the camera. This footage will have the letters written on the glass aa inverted horizontally. Then this footage is flipped horizontally by editing software.
I was wondering that. If I had to do something like that, I would place him behind a glass wall, turned to the camera which on the other side of the glass. Then mirror the recorded video. But maybe they do it differently
I think he did write in a reverse way!
That's it folks.. It's happening.. I'm scared about what this technology can do
Hold up... Am I watching a video on LLMs from the same guy that I've learned about home brewing from?!?!?
Is he right or left handed
So this is what Ole is upto after leaving Man Utd
I was a little scared that this would be a blockchain thing.
Scared ? Lol
Scared?? Weak soul spotted
It could be used for that. Get that schmoney.
honestly, I dont really like lang chain, I'd rather write the code myself to do this because its pretty trivial. seems more like something for non-techinical people who arent in tech
4000 laoff today. are you affected.
why?why?why?Why IBM know everything but lose AI age?
IBM will soon shock everyone 😅. You wait for it 🙂
Everything, what could be done with langchain, better to be done without langchain. Langchain is everything bad from huggingface, without anything good, except false promise to make AI go brrrr with few lines of code. Okay, it will go brrrrr. But slowly. And when you'll need something at least a bit uncommon, you'll find, that everything is obfuscated behind several levels of unnecessary abstractions. And you'll need to rewrite everything.
subscribed
Way to break down the chain. Well done.