LangChain4J - use the power of LLMs in Java!
HTML-код
- Опубликовано: 6 мар 2024
- An introduction to LangChain4J : what it is, why use it, with 7 short demos showing its main use cases in Java. We'll use Azure OpenAI and MistralAI, Dalle-3, GPT-4 and Mistral 7B, learn how to use embeddings and vector databases, and how to use the RAG pattern.
The demos use a Spring Boot application that is easy to understand and re-use in your own projects. It is available at github.com/jdubois/2024-LangC... 
Наука
Sir spectacular work I loved to see this amazing tutorial....
Coool, thanks for the great tutorials!
Thanks Julien
thanks Julien, great video!
Thank you for this presentation.
I am glad you like it!
Tanks Julien, it was enlightening!
Thank you for sharing !
Great video
god damn that accent is killing me. man I love that accent.
Is it already possible to build multi-agent applications in Spring AI like Langchain Multi-Agents, DSPy or Autogen? Are there any good examples out there?
The short answer is no, not yet. The longer answer: It's technically possible, but at ths point in time only insofar as you create a lot of the constructs+mechanisms for agent interaction/integration yourself. Not speaking for the Spring AI team here, but the general path of evolution for Spring projects is to start with a kernel of utility/abstraction and build out from there. It's still very early days for Spring AI (and for AI in general of course, but even more so for Java+AI of any flavor) so I suspect there will be constructs at some point that fall within or adjacent to the Spring AI project. For reference, consider Spring Cloud from a few years back: it built upon and around Spring Boot apps to provide what I like to refer to as cloud scaffolding, enabling Boot apps to work seamlessly *together*. I suspect this will be the natural evolution in this case as well.
Again, no insider knowledge, just an eye for where the ball is going. I think. YMMV. :)
@@MarkHeckler Thanks for the very detailed answer. And yes, I know it's possible to exactly start as you described and I also assume that it will be matter of time until Sping AI offers a set of more mature tools for agents. Many enterprises are currently focused on reducing workforce, so they will test it as soon as possible when it's available but it's also clear that Spring AI needs some time for newer tools and the Python related AI stuff will always be some months ahead. Anyway, I have my eyes on it because it will affect my job for sure, as an employee and as a developer 🙂
@@returncode0000 Agreed on all counts. We're all eagerly keeping an eye on things to see how it all develops (no pun intended). Here's to the future!
@@MarkHeckler Yes! Hope to have a little advantage because we (my employer) has a very big partnership with Azure and Open AI :)