Microsoft Autogen Studio 2 - How to run an army of agents

แชร์
ฝัง
  • เผยแพร่เมื่อ 15 ม.ค. 2025

ความคิดเห็น • 41

  • @DonaAI
    @DonaAI  11 หลายเดือนก่อน +3

    Let me know here what workflow you want to streamline with your army of agents 👇🏻✨

    • @fredericherrera
      @fredericherrera 11 หลายเดือนก่อน

      Same agents but with Twitter comments! Good video thank you

    • @Vigilence
      @Vigilence 11 หลายเดือนก่อน

      Search websites for the first few pages of products for a specific keyword, organize the list in a visually pleasing way along with a picture of the product, url link, product title, price, and average review score.

    • @lowkeylyesmith
      @lowkeylyesmith 10 หลายเดือนก่อน

      Hi and have a nice day,
      I am trying to realize a project that connects several OpenSource LLM's with an agent framework. Each LLM is supposed to solve a specific task. Unfortunately, I can't describe it as well as I would like to, but I'll give it a try.
      It is about receiving, unpacking and analyzing a PST file. There should be a separate trained LLM for each area of the analysis. For example, an LLM that graphically displays the connections of the suspect to others, one that is trained with different languages to search in all possible languages, .... so the whole thing is briefly outlined.
      Since I am still relatively new to this area, I would be very happy to receive tips and tricks.
      Thank you and best regards
      René

  • @evelynshi4475
    @evelynshi4475 11 หลายเดือนก่อน +3

    Video quality is getting better and better 🤗

  • @EvelynShi-l1b
    @EvelynShi-l1b 11 หลายเดือนก่อน +3

    I was looking for tutorials for this so thanks man!

  • @warezit
    @warezit 11 หลายเดือนก่อน +1

    🎯 Key Takeaways for quick navigation:
    00:00 🤖 *Introduction to Autogen Studio*
    - Overview of Autogen Studio's capabilities to run multiple AI agents with specific skills.
    - Each agent utilizes different machine learning models for assigned Python functions.
    00:26 🛠️ *Setup and Usage Guide*
    - Step-by-step guide on setting up and using Autogen Studio's latest version.
    - No coding knowledge required for running workflows using local machine learning models.
    01:22 📚 *Installation and Model Configuration*
    - How to install Autogen Studio and configure different models including GPT-4.
    - Detailed explanation on adding new models and testing their setup.
    02:21 🖥️ *Local and Remote LLMs*
    - Differences between using local machine learning models and remote models like GP4 preview.
    - Insights on adding models to Autogen Studio through LM Studio.
    03:23 ⌨️ *Creating Skills for Agents*
    - Process of adding new skills to agents through simple Python functions.
    - Example provided on creating a function for fetching TH-cam comments.
    04:18 🎭 *Agent Creation and Configuration*
    - Steps to create agents with specific roles and skills within Autogen Studio.
    - Explanation on naming agents and assigning them predefined skills.
    05:42 🔄 *Creating Workflows with Agents*
    - Workflow creation logic combining multiple agents for a series of actions.
    - Understanding of how user proxy agents manage the flow of tasks between agents.
    07:08 🛠️ *Workflow Step-by-Step Guide*
    - Detailed guide on setting up a workflow in Autogen Studio to generate TH-cam video content.
    - Includes step-by-step instructions from capturing TH-cam comments to generating content ideas.
    08:02 📈 *Workflow Execution Results*
    - Demonstration of the final result after running the created workflow in Autogen Studio.
    - Insights gained from the comments captured and content ideas generated.
    09:29 🚀 *Conclusion and Further Applications*
    - Recap of capabilities showcased in the tutorial and potential applications.
    - Invitation for viewers to explore developing their own agents and workflows.
    Made with HARPA AI

    • @warezit
      @warezit 11 หลายเดือนก่อน

  • @samatkadyrov5779
    @samatkadyrov5779 11 หลายเดือนก่อน +2

    Woah man, this is great! Please make more such tutorials!

  • @ecofriend93
    @ecofriend93 11 หลายเดือนก่อน +1

    Thanks for the thorough explanation!

  • @sebastiencailhol8020
    @sebastiencailhol8020 11 หลายเดือนก่อน +2

    Great video, many thanks

  • @avi7278
    @avi7278 11 หลายเดือนก่อน +1

    Merci Jean Jacques

  • @landeregia
    @landeregia 11 หลายเดือนก่อน +2

    Great video! Thanks a lot Dona.

  • @this.theexperience
    @this.theexperience 9 หลายเดือนก่อน +1

    I keep running into an issue where the agent doesn’t follow the prompt instructions. When I ask it for its prompt it repeats back the initial prompt or a prompt from a prior session even though this is a new session. Did you have this issue? The problem this presents is all the agents responding to all the requests.

    • @chirantanbhardwaj9248
      @chirantanbhardwaj9248 8 หลายเดือนก่อน

      I'm having the same issue man, if is soo frustrating. That's why I switched back to autogen

  • @StephenRayner
    @StephenRayner 11 หลายเดือนก่อน

    Nice mate, use GPT everyday extensively to right high quality code (13 year experience). I found this video very useful and well explained.

    • @StephenRayner
      @StephenRayner 11 หลายเดือนก่อน

      Oh length of video is good. Glad it’s not longer than 10 min

  • @adambenari3944
    @adambenari3944 11 หลายเดือนก่อน +2

    Thanks Dona!

  • @JuanRamirez-di9bl
    @JuanRamirez-di9bl 10 หลายเดือนก่อน +1

    Let's say you want a local multi agent workflow, would you have to load all models to memory if they are different models? I am assuming yes, what if I want to use a the same model with multi agents, do I need to load the same model to memory several times? Or how does that work?

    • @DonaAI
      @DonaAI  10 หลายเดือนก่อน +1

      If you want different models running locally you will have to run each model in a different port (or server) and have as many APIs endpoints as models you want to run. They will need to all be loaded in memory and/or gpu.
      If you use the same model, you will have one API endpoint that each agent would call (sequentially) so you just need to load it once in memory.
      If multiple users are using the workflow or multiple calls are done in parallel, I think it would depend on the architecture of the LLM. Not sure about that case but I will research about that. 🤔

    • @JuanRamirez-di9bl
      @JuanRamirez-di9bl 10 หลายเดือนก่อน

      @@DonaAI interesting! Thank you for your answer!

  • @TheQuantumCapitalist
    @TheQuantumCapitalist 6 หลายเดือนก่อน

    The question I've yet to see asked or answered is, "how do you create login" info and save the data you store in the UI? When I access the platform, after adding agents and skills, its as if I never did anything. Everything is back to default.

  • @shivangchaturvedi237
    @shivangchaturvedi237 11 หลายเดือนก่อน

    Yeyyyyy... Lets go Dona!!!!🚀🚀

  • @snuwan
    @snuwan 10 หลายเดือนก่อน

    Is there a specific format for the skills or way to instruct the agent to use the skill. Also I think you can not run skills with open source models. I found out that only Open Ai models can run these skills. Locally hosted models would not run them. Any idea ?

  • @MrTactics26
    @MrTactics26 8 หลายเดือนก่อน

    Is there a way to save the 8081 environment? Every I set it up it times out after like 10mins. I want to be able to come back later in the day and keep working on it. Thanks

  • @DigitalForest0
    @DigitalForest0 11 หลายเดือนก่อน

    thank you, is there a way/idea i could use python and autogen to automate still image animations like i do on my channel? or at least partially automate?
    this would be revolutionar!

  • @StephenRayner
    @StephenRayner 11 หลายเดือนก่อน

    Question Twitter/X LLM has a tree branch like UI. Do you know of any open source project that have this? With how I use LLMs this would be very useful but I cba to build it myself.

  • @RoaldReurink
    @RoaldReurink 11 หลายเดือนก่อน

    Great thank you! I've got it almost working ... keep getting this reply "As an AI language model, I don't have direct access to external databases or APIs, including TH-cam's API, to fetch comments or perform any actions on TH-cam."
    I've the API key, also added it in script to test, but nothing :( What am I doing wrong?

  • @unimposings
    @unimposings 10 หลายเดือนก่อน

    is it become better now, i mean it was really buggy 3 weeks ago.

  • @HideBuz
    @HideBuz 11 หลายเดือนก่อน

    Never heard "snake case" before, we just call them underscore. Good content! Which chatgpt model do you think is most cost effective?

    • @PablumMcDump
      @PablumMcDump 11 หลายเดือนก่อน

      snake_case is lowercase with underscores, to differentiate it from camelCase which puts words together but capitalizes the first letter of each word but the first.

  • @mack6807
    @mack6807 10 หลายเดือนก่อน

    pip install -U autogenstudio
    This fixed my Missing Column in Database Error and now I can add agents.

  • @bekhzodortikov421
    @bekhzodortikov421 10 หลายเดือนก่อน

    Hey, bro! Can you help me this this prblem:
    AGENTS
    TERMINATE
    Agent Messages (3 messages) | 1 sec
    userproxy ( to adminstrative_chat )
    hey
    adminstrative_chat ( to userproxy )
    Hello! How can I assist you today?
    userproxy ( to adminstrative_chat )
    TERMINATE
    I'm getting several times those TERMINATE massega how can i hide it? Cause I see you dont getting this message(

    • @this.theexperience
      @this.theexperience 9 หลายเดือนก่อน

      Ensure you don’t have terminate as the default response.

  • @user-wr4yl7tx3w
    @user-wr4yl7tx3w 11 หลายเดือนก่อน +1

    fonts need to be bigger i think

    • @DonaAI
      @DonaAI  11 หลายเดือนก่อน

      Good point! Thanks for the feedback!

  • @aleclippe6213
    @aleclippe6213 11 หลายเดือนก่อน

    How much for 3 hours of your time? Serious question

    • @DonaAI
      @DonaAI  11 หลายเดือนก่อน

      Feel free to drop me an email if you need some help on your project 😉

  • @countermeasuresecurityengi9719
    @countermeasuresecurityengi9719 10 หลายเดือนก่อน

    My PC is an old HP Z620 server only has VTx, can I still install Autogen Studio AND which LLM should I be able to install locally if I can use this old server which has 64GB RAM? Thx for the video!