CrewAI is Better than AutoGEN ?? Use with Ollama Openhermes !

In this video, we are going to explore CrewAI which is a Cutting-edge framework for orchestrating role-playing, autonomous AI agents.

CrewAI's Advantage: CrewAI is built with production in mind. It offers the flexibility of Autogen's conversational agents and the structured process approach of ChatDev, but without the rigidity. CrewAI's processes are designed to be dynamic and adaptable, fitting seamlessly into both development and production workflows.

Let’s do this!
Join the AI Revolution!

#ai #crewAI #autogen #joinpatreon #assistantAPIs #GPTs #llm_selector #auto_llm_selector #localllms #github #streamlit #langchain #qstar #openai #ollama #webui #github #python #llm #largelanguagemodels

CHANNEL LINKS:
🤠Join this channel to get access to perks:

If you have any questions, comments or suggestions, feel free to comment below.
🔔 Don't forget to hit the bell icon to stay updated on our latest innovations and exciting developments in the world of AI!

[HistoryIsAbsurd]

Hey thanks alot it took me all day (im new still to coding) but with your help im now up and running locally. Worth the sub!

[RainbowSixIntel]

This is SO much easier to run and install than autogen. Thank you very much for this!

[fslbceb]

I like this approach; its quite informative to something I'm developing. Very much appreciated.

[mainajnabee]

Great video and i'm curious to work on it! Thank you!

[EmilioGagliardi]

I've been seeing some buz about Ollama but haven't looked into it. How does it perform on structured output? Is It possible to create fine-tuned instances of Ollama? I'm using openai to extract graph entities and its quite costly and I'm wondering if Ollama is good enough yet or if it can be tuned. Helpful video, thanks.

[somare]

Thank you for great work, just a question, never really worked with Lang chain, does it have native browsing support? how does these agents do research? (I'm assuming they browse online.)

[DrFukuro]

Thank you for your insights! But have you checked the output of the second bot? Topic failed. It described the task in detail, but did not finish it. I would be interesting how a guy like you (with your nickname) would tackle this problem. And where is/could be the correction involved like the user-proxy in autogen?

[samfakhoury]

Can you teach us how to make CrewAI work with Gemini 1.5 ?

[Justin-C]

I'm pretty sure your example at the end - 11:41 - didn't even work without the Action set up. It seems like it just resorted to an answer cooked up by the LLM without actually fetching new articles like it's supposed to.. look where it says "...is not a valid tool, try one of []"

[terrencekunstek]

Does crew use tools and perform actions? i.e. Can you have it scrape and summarize website content to include in it's research?

[cachipum]

Man CrewAi is no way better than autogen, is just a simpler maybe more practical approach for some use cases, but autogen is definitely more flexible and powerful, crewAI idea seems good but it is in a really early stage, I was trying crewai the whole weekend to do different things... it is specially slower than autogen for some tasks

[SaveTheDoctor-flhn]

I assume since you are using langchain and calling the LLMs from the code, you could in theory call dedicated different LLMs for each agent?

[igormedeiros]

hi, I have my llm running on a local server, so I need to create this object informing the host/url:
ollama_openhermes = Ollama(model='openhermes')
Do you know how can I do it?
Congrats by your content!

[rs]

Thanks for all the great content you share wuth us! CrewAi has released version 2 with hierarchial process --> could you make a video with crewAi using memGPT (big bonus points if it also uses open interperter!)?

[engkc]

Did it actually work? Geeting error "ImportError: cannot import name 'Agent' from partially initialized module 'crewai' (most likely due to a circular import)"

[theabhijitdn]

Amazing
. Kindly do the same with Gemini

[mackroscopik]

Would it work using a local LLM and LM Studio?

[googleyoutubechannel]

So I was just trying out 'crewAI' as it's gotten so much traction on youtube and it's well... pointless?
I've run the examples here and the frameworks specific ones, and while they're nice and do what they say, they're all just a linear list of prompts that pass output to the next prompt... wheeee.
(and are impossible to debug, everything happens via a single 'run' function)

Why has it gotten so much attention, it seems to be just an abstraction that is basically just a list and switch statement, one that then hides everything behind abstractions that make it nearly impossible to debug as easily as just using a list of prompts?
There's always a trade-off between learning the specifics of these abstraction layers and customizability, but in this case the choice seems super easy, this is a complex abstraction layer that hides critical info and forces a very opaque specific framework.... that makes things like a few lines of code easier... so completely pointless?

The best thing it has going for it is that the single dev seems to have spent a lot of time on marketing it and docs, which is nice I guess?

[juniorpowell]

Hey I need your help my brother do you have a social media where we can talk