What to understand more about LangStream's event driven architecture and how it helps when building LLM applications?
#genai#llm#streamprocessing#kafka
"Unpacking LangStream's Event-Driven Architecture"
https://t.co/FWXhF8XFzI
Check out this great blog post by @DierufDavid detailing all the great new features of LangStream 0.4, including:
* Stateful agents
* Apache Camel as a source
* "Service" components
* "Service" gateways
https://t.co/ETLHlVkpHt
You can now develop with LangStream using open-source models like Llama2 on your laptop using @Ollama_ai.
Check out this new blog post from @eolivelli explaining how to do it.
https://t.co/Iyoyls9nHm
Streamline your data with LangStream's latest guide by @eolivelli on scalable vectorization pipelines! 🚀
From S3 buckets to vector databases, master the flow of information with ease. Dive into the full pipeline process here!
#LLM#GenAI#Vectorization
https://t.co/4mlBmR1c9R
Yesterday we've had another @langstream_ai meetup in Bologna with the local Java User Group and other folks!
That was fun, and we had great discussions about #GenerativeAI !
@eolivelli
Want a step-by-step tutorial on how to build a web crawler chatbot using LangStream?
Check out this great new video by @DierufDavid from the Thinking Machine YouTube channel.
https://t.co/tBiPdfL3MT
In the LangStream project, we are always trying to make it easier for you to develop cool LLM applications. We recently added a UI for testing and debugging your LangStream apps. Check out the video for a quick demo:
https://t.co/m40KK3L9lE
1/8 🧵 Building a chatbot that can answer questions about private data? Let's talk about using Retrieval Augmented Generation (RAG) pattern. This pattern helps large language models (LLMs) provide fresh answers about topics they weren't trained on. Here's a detailed breakdown.👇
#RAG: going from the theory to real production-ready applications is not easy, but if you use @langstream_ai that’s another story
New blog post that explains step by step how to do it https://t.co/ynaByYtaCh
🔔 We just released LangStream 0.2.0 with new cool features:
* Support for FLARE pattern: FLARE is an emerging pattern, similar to RAG, that iteratively uses a prediction of the upcoming sentence to anticipate future content, which is then utilized as a query to retrieve relevant documents to regenerate the sentence if it contains low-confidence tokens. Read more in the paper (https://t.co/rRwMcqZjk7).
* New agent ‘dispatch’: you can implement message routing easily by defining conditions on the record.
* New agent ‘http-request’: run HTTP requests to enrich your data model from any public service or call webhooks when an event happens on the pipeline.
* New agent ‘azure-blob-storage-source’: read and process files from Azure Blob Storage containers.
* New UI for apps monitoring and testing: you can now spin up a local UI web server to monitor and test your application in a friendly way.
* CLI support for Windows: CLI is now installable on Windows.
* Export your application as Mermaid graph: run langstream apps get myapp -o mermaid to generate a MermaidJS (https://t.co/Oqv0D251vV) graph that shows you the entire application workflow.
Checkout the full release notes: https://t.co/MHlhDrTKvc
Please add your ⭐ to spread the project!
@mohinishbasha@Meetup Hi Mohinish. We'd love to compare notes on your demo chat bot application and what we are doing in the LangStream project. We might
📢Checkout my new blog post if you're delving into the world of streaming SQL workloads, seeking to slash data/ML infrastructure costs by millions, or frustrated by the poor optimization coordination across processing and storage systems in your data org! https://t.co/65IvwuTnmr