Qwen-Agent: Build Autonomous Agents with The Best Open Weight Model
In this video, I dive into Alibaba’s latest Qwen-2 model, the best open weight model available. We explore function calling and agentic workflows using Qwen-2’s powerful features. The model family ranges from 0.5 billion to 72 billion parameters, with support for up to 128,000 tokens and multiple languages, including Middle Eastern and Southeastern languages. I’ll show you how to perform function calling and create custom agents with the Qwen-2 agent framework.
💻 Massed Compute VM: https://bit.ly/localGPT (use Code: PromptEngineering for 50% off).
💻 RAG Beyond Basics Course:
https://prompt-s-site.thinkific.com/courses/rag
🦾 Discord: https://discord.com/invite/t4eYQRUcXB
☕ Buy me a Coffee: https://ko-fi.com/promptengineering
|🔴 Patreon: https://www.patreon.com/PromptEngineering
💼Consulting: https://calendly.com/engineerprompt/consulting-call
📧 Business Contact: engineerprompt@gmail.com
Become Member: http://tinyurl.com/y5h28s6h
Signup for Newsletter, localgpt:
INSTRUCTION:
conda create -n qwen python=3.10
conda activate qwen
LINKS:
Qwen-2: https://qwenlm.github.io/blog/qwen2/
Qwen-Agent: https://github.com/QwenLM/Qwen-Agent
Function Calling Example: https://qwen.readthedocs.io/en/latest/framework/function_call.html
Agent Usage Example: https://qwen.readthedocs.io/en/latest/framework/qwen_agent.html
Gemini Flash Agents: https://youtu.be/A20wzlC7Q-c
TIMESTAMPS:
00:00 Introduction to Quen2 Models
01:51 Difference between function calling and Agents
04:31 Setting Up and Running Quen2 Locally
07:31 Function Calling with Quen2: A Practical Example
11:54 Creating Custom Agents with Quen2
17:04 Impact of Quantization on Model Performance
All Interesting Videos:
Everything LangChain: https://www.youtube.com/playlist?list=PLVEEucA9MYhOu89CX8H3MBZqayTbcCTMr
Everything LLM: https://youtube.com/playlist?list=PLVEEucA9MYhNF5-zeb4Iw2Nl1OKTH-Txw
Everything Midjourney: https://youtube.com/playlist?list=PLVEEucA9MYhMdrdHZtFeEebl20LPkaSmw
AI Image Generation: https://youtube.com/playlist?list=PLVEEucA9MYhPVgYazU5hx6emMXtargd4z
ubuntu download
i've found other models performing much better in an agentic workflow across many use cases, many times the results are nonsense from qwen2, so it needs to be filtered out, but the 128k context length is nice.
Can you explain how to make it english
I have random question regarding massedcompute. How many hours do you typically run the LLM for? 1 hour? more? less? And how much do you spread it across? 1 week (assuming you're testing it or you're building something with it)? Also what's the overall cost?
Glad you clarified the definition of an agent! Many mix it up with calling multiple consecutive LLM calls which is “pipeline” and not agent. An agent needs autonomicity to plan, think, and decide. Also, I'm creating a COLAB to challenge the model's function calling ability. I'll share it soon for you to use and review.
Another thing I wonder if their AWQ quantization is 4-bit or 8-bit. The table suggests 4-bit because AWQ preserves accuracy better and scores better than other 4-bit methods. But it scores lower compared to 8-bit, indicating they use 4-bit. I think AWQ 8bits is the best for most of cases.
On ollama we have Q4. How do pull another quantzed model?
Gracias
How can u run a small model in a phone
so even iQ2-XXS of 70B is still far better than fp16 7B?
Very informative, just curious why not langgraph to manage this.