Відео

@@franknillard Thanks - bring it on!

@@ParthShukla-o3t yes - check out the video description 👍👍

@@thomasderidder608 They can split into as many paths as you like. You can also creat an unknown amount of edges to complete in parallel with the 'send' class

@@ahmedms6947 Glad you liked it! Yes that model should be no problem. What line does the error relate to?

@@rajashekarakula3991 Thanks! If I do a follow up, what is it that you want to know?

I tried that before and it was directing me somewhere that wasnt their studio... This time it worked though, so maybe some session funniness in the browser

🙏 thanks

@mariuszsiera thanks for the feedback. I have used 2/3 of them and wanted to see how the other shaped up. I considered something like that but as I was creating this video it felt like comparing apples to nuts. They are all quite different and have different problems that they would work better at solving.

I would consider a tiny docker instance so that I have full control over the communication & variable management. This article gives a walkthrough: anukriti-ranjan.medium.com/building-a-sandboxed-environment-for-ai-generated-code-execution-e1351301268a

@@WW_AI_Adventures Thank you so much!

@@andrew.derevo I certainly got that feeling about Autogen. CrewAI looks ok as a framework but I think only if what you are trying to achieve is task based and well defined.

@ we use langchain with lightllm as unified api proxy provider, works really good for now, langchain sometimes feels overloaded in some cases but definitely do the job. will check crewai again, probably something changed over time.

@@WW_AI_AdventuresCreeAI has limited documentation path to true production deployment at scale in the Cloud compared to LangGraph that offer multiple options: SaaS, Cloud (container) or even local dev! So for me CrewAI is the weakest in this area

@@reserseAI looks good tbh! Another good out of the box option. A templated API like this would be good if you could then build the templates with an orchestrator like LangGraph

Ah that's good to know 🙏

Thanks! 🙏

Oh yeah? I'm not so sure..

Haha maybe you’re right, but I find it nice. Not to many abstractions and easy to get insight with log fire. Have you tried it?

@alexwoxst Not yet - I had a look and it does look like a good API with simple abstractions, I will be taking a more detailed look though!

Pydantic ai is a revolution for me in making commercial grade ai apps. It's easy to swap out llms as requirements change so many other benefits

@Jobeyhshxgs Interesting, thanks for letting me know.

I like the fact that it integrates with HuggingFace models, so it could be useful for anything that you want to do quickly or for a proof of concept, as integrating with on device models can be a pain! It looks simple, so it depends what you want.

Hiya! Thanks for the feedback. Do you have an error you can share?

@@WW_AI_Adventures thanks for the prompt response! It seems the error occurs when microsoft_to_neo4j.py tries to read the columns 'name' and 'description_embedding' from create_final_entities.parquet, and 'rank' from create_final_relationships.parquet, as these columns are missing from these files. However, the rest of the columns are present. Don't know what is causing it configuration, models or something else.

@MuhammadSadiq-k3q it could be that Microsoft has changed their format for the parquet files. Try inspecting the files manually to see if the names have changed

@@webclubco thanks 🙏

Thanks for the suggestion 🙏

Thanks - Glad you liked it!

Glad you enjoyed it!

@@GauravWankhede-x5q Yes you could absolutely use that model ai.google.dev/gemini-api/docs/models/gemini-v2#improved-tool-use You might need to have a play around with the prompt & tool arguments though to make it work well.

​@@WW_AI_AdventuresThank You sir for the reply, My Planning is to use Gemini based System which will Automate the Data Science project. And I am going to use FastAPI, hence I can Integrate those API with any Tech Stacks

@@rtrvr-ai Thanks for the tip, will check it out

@@michaelmalone7614 Thanks! What did you like about it?

@@WW_AI_Adventures You're welcome. What did I like about it? Professional looking presentation, well thought out structure (outline plan for tutorial, explain concept then demonstrate with code) and best of all, a complete absence of waffle, succinct, efficient, straight to the point. Far too many of these types of tutorials contain meandering diatribes and fluff. Anyway, hope that helps.

​@@michaelmalone7614 Yes thank you much appreciated!

@@syedhaideralizaidi1828 Thanks Syed! What applications are you working with? Let me know your real world problems and I'll consider doing a vid :)

@@BhuvanOberoi-b9v No major limitations, Claude supports tool calling in the same way OpenAI does. If you want to use LangChain and the @tool decorator then that's fine as well!

Hiya! Thanks for commenting - Yes I used the same LLM for both. GPT-4o. To be honest this isn't a strict benchmark, just my exploration of the two together. Naïve RAG will simply never be able to include all of the text data from a large corpus in its context window. GraphRAG gets around this by precomputing summaries ahead of time - so it will always have an advantage, at the cost of this ahead of time summarisation which may not be possible for an incredibly large corpus!

@@jackbauer322 thanks for the suggestion. LightRag does look good. However, if you want a true global summary of all of your text, then I don't know if it will be able to do as good a job!

Thanks for the suggestion!

Thanks 🙏

@@arthuraquino8356 Thanks! I'm interested in Text2SQL, but there are lots of structural problems with analytics that an agent won't be able to solve. Like stakeholders not knowing exactly what they want to see & data often being very messy and only the analysts knowing what data is reliable! Do you have a particular problem within text to SQL?

@@WW_AI_Adventures I have been trying to use pgvector with hybrid search to make specific queries and then have the agent use the SQL query tool because the select query needs to be exact. For example, what is the stock of the product "chair model jess", where if I misspell a letter and the model does not have the nuance, it will confirm the error in the select. That is why I am trying to validate the term before using the hybrid search and then make a query. But I don't know if this would be a better way.

@@iwswordpress thanks! Glad you liked it

@@riteshbhadana I'm sure I can manage that!

@@WW_AI_Adventures thanks for your reply pl do it fast

So glad you liked it!

@@WW_AI_Adventures 👍🏾

@@JCastillo I haven't got a tutorial with java script (yet!) but the documentation is here langchain-ai.github.io/langgraphjs/

Thanks, I'm glad you liked it! Anything else you'd like to see next? Any concepts you're struggling with atm?

@@WW_AI_Adventures Thanks for asking. I'm exploring LangChain/LangGraph on how to use them to rapidly create a basic chatbot for use cases such as customer support or answering user questions on a college website.

@@Schimiling No problem - thanks for tuning in!

@@sagartamang0000 Hiya, Thanks for getting in touch. I would be surprised if you are forced to use a specific LLM. You could still support function calling with any LLM, but this would likely lead to poor results. If you are constrained by needing to use open source then I would consider Hermes or LLama 3 which have both been trained to support function calling. If you are needing open source AND on device, then I would consider TinyAgent from Berkley - bair.berkeley.edu/blog/2024/05/29/tiny-agent/ Even if these models don't support the 'bind_tools' function from LangChain, you could still use them for function calling, you would just have to look up the specifics of the model as to how you encode tools and extract selected tool calls from the LLM output. Sorry I can't be more specific, it depends on the LLM! Best, Will W

@WW_AI_Adventures my my, that was an in-depth response. I am so thankful to find you! Also I am enjoying the content you've posted, thank you again!!

Hiya! I have a tutorial series that goes into more detail, so feel free to check that out. ua-cam.com/play/PLojmSSBcl4H9Cj8AnaG7xMPgU5zkDf2iZ.html&si=a1frPpE7rEB9HJhh I haven't got anything super complicated currently relating to langgraph state management. Let me know if you find anything!

@@WW_AI_Adventures Thank you very much!

@@siddharthshukla3557 Hiya! What I have done in the past is to periodically summarise all the messages past a certain point. For instance when you reach over 20 messages, every 5 messages, you can summarise the messages over 20 and pass in the previous summary as well. This way you alway have all the important information in the context without the bill increasing non stop. I hope that helps, if not then please get in touch on my discord community discord.gg/VgsdC8nk where id be happy to help!

Thank you! Will do 😃

Thanks. I've been thinking about something like this - different agentic architectures in langGraph - or did you have something else in mind?

@@WW_AI_Adventuresyes please cover it . Some basic to advance service. How to customise it how to create node & diff workflow