TypeError: LocalStateManager.__init__() got an unexpected keyword argument 'agent_config'

[mongolu]

Describe the bug I'm trying to run the code from autogen from this file: memgpt/autogen/examples/memgpt_coder_autogen.ipynb A clear and concise description of what the bug is. In file memgpt/autogen/memgpt_agent.py at line 170, i think it should've been "agent_state": agent_config, .

LOGS:

Traceback (most recent call last):
  File "/LLMmodels/autogen/ytest.py", line 75, in <module>
    coder = create_memgpt_autogen_agent_from_config(
            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/MemGPT/memgpt/autogen/memgpt_agent.py", line 95, in create_memgpt_autogen_agent_from_config
    autogen_memgpt_agent = create_autogen_memgpt_agent(
                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/MemGPT/memgpt/autogen/memgpt_agent.py", line 172, in create_autogen_memgpt_agent
    persistence_manager = LocalStateManager(**persistence_manager_kwargs) if persistence_manager is None else persistence_manager
                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
TypeError: LocalStateManager.__init__() got an unexpected keyword argument 'agent_config'

Please describe your setup

• [x] MemGPT version
  • What is the output of memgpt version? (eg "0.2.4") 0.2.12
• [x] How did you install memgpt?
  • pip install pymemgpt? pip install pymemgpt-nightly? git clone? git clone
• [x] Describe your setup
  • What's your OS (Windows/MacOS/Linux)? Linux
  • How are you running memgpt? (cmd.exe/Powershell/Anaconda Shell/Terminal) terminal

Screenshots If applicable, add screenshots to help explain your problem.

Additional context Add any other context about the problem here.

---

If you're not using OpenAI, please provide additional information on your local LLM setup:

Local LLM details

If you are trying to run MemGPT with local LLMs, please provide the following information:

• [ ] The exact model you're trying to use (e.g. dolphin-2.1-mistral-7b.Q6_K.gguf)
• [ ] The local LLM backend you are using (web UI? LM Studio?)
• [ ] Your hardware for the local LLM backend (local computer? operating system? remote RunPod?)

[mongolu]

After modifying from "agent_config" into "agent_state", other error pops out:

Traceback (most recent call last):
  File "/LLMmodels/autogen/ytest.py", line 75, in <module>
    coder = create_memgpt_autogen_agent_from_config(
            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/MemGPT/memgpt/autogen/memgpt_agent.py", line 95, in create_memgpt_autogen_agent_from_config
    autogen_memgpt_agent = create_autogen_memgpt_agent(
                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/MemGPT/memgpt/autogen/memgpt_agent.py", line 172, in create_autogen_memgpt_agent
    persistence_manager = LocalStateManager(**persistence_manager_kwargs) if persistence_manager is None else persistence_manager
                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/MemGPT/memgpt/persistence_manager.py", line 54, in __init__
    self.archival_memory = EmbeddingArchivalMemory(agent_state)
                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/opt/MemGPT/memgpt/memory.py", line 366, in __init__
    self.embed_model = embedding_model(agent_state.embedding_config)
                                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AttributeError: 'AgentConfig' object has no attribute 'embedding_config'. Did you mean: 'embedding_dim'?

[cpacker]

Hey @mongolu , sorry for the delay! This is basically a version issue - the latest code on this github repo is for the eventual 0.2.12 release, but the AutoGen part of the repo has not been updated yet so it won't work until #787 is finished.

If you want to use AutoGen, you should use 0.2.11. You can do this by grabbing the pip package (pip install pymemgpt or pip install pymemgpt==0.2.11), or by cloning the github repo then doing git checkout 0.2.11).