How to start the server

[mrobinson-opi]

Hi thanks for all your work. I am trying to get started with this on an RK3588 device. I have ollama installed and cloned the repo and installed as much as I could figure out. but I am getting this error and I'm not sure if I am even trying to do it right. Thank you.

(.venv) mrobinson@orangepi5pro:/MemOS$ poetry run python -m memos.api_server
/home/mrobinson/MemOS/.venv/lib/python3.12/site-packages/qdrant_client/http/models/models.py:758: SyntaxWarning: invalid escape sequence '&'
description="Check that the field is empty, alternative syntax for is_empty: \"field_name\"",
/home/mrobinson/MemOS/.venv/lib/python3.12/site-packages/qdrant_client/http/models/models.py:762: SyntaxWarning: invalid escape sequence '&'
description="Check that the field is null, alternative syntax for is_null: \"field_name\"",
/home/mrobinson/MemOS/.venv/bin/python: No module named memos.api_server
(.venv) mrobinson@orangepi5pro:/MemOS$ poetry run python -m memos.server.main
/home/mrobinson/MemOS/.venv/bin/python: Error while finding module specification for 'memos.server.main' (ModuleNotFoundError: No module named 'memos.server')
(.venv) mrobinson@orangepi5pro:~/MemOS$

[Ki-Seki]

Simply run make serve or the equivalent command poetry run uvicorn memos.api.start_api:app in the MemOS folder to start the server.

[mrobinson-opi]

Thanks for the responce!
Another questions how do I change the port number?

[Ki-Seki]

Sorry, I didn't understand your question. Can you express it in a different way?

[mrobinson-opi]

How to use MemOS with OLLAMA?

[mrobinson-opi]

I feel like I am getting closer
(.venv) mrobinson@orangepi5pro:~/MemOS$ python3 test_v2.py
/home/mrobinson/MemOS/.venv/lib/python3.12/site-packages/ollama/_types.py:81: PydanticDeprecatedSince211: Accessing the 'model_fields' attribute on the instance is deprecated. Instead, you should access this attribute from the model class. Deprecated in Pydantic V2.11 to be removed in V3.0.
if key in self.model_fields:
MemOS client initialized and connected to Ollama!
An error occurred: shapes (0,1536) and (768,) not aligned: 1536 (dim 1) != 768 (dim 0)
Please ensure your Ollama server is running on http://localhost:11434.
You can verify Ollama status by visiting http://localhost:11434/ in a browser on your Orange Pi.
Also, double-check that 'llama3:latest' and 'nomic-embed-text' are present when you run 'ollama list'.

[mrobinson-opi]

/home/mrobinson/MemOS/.venv/lib/python3.12/site-packages/ollama/_types.py:81: PydanticDeprecatedSince211: Accessing the 'model_fields' attribute on the instance is deprecated. Instead, you should access this attribute from the model class. Deprecated in Pydantic V2.11 to be removed in V3.0.
if key in self.model_fields:
MemOS client initialized and connected to Ollama!

[Ki-Seki]

Thanks for all your work getting MemOS and Ollama up and running — looks like you're very close!

From the error message shapes (0,1536) and (768,) not aligned, it seems there's a mismatch between the embedding output from Ollama and what MemOS expects. A few things you can check:

• ✅ Make sure the embedding model (nomic-embed-text) you're using produces 768-dimensional vectors — or adjust MemOS to handle the actual output size if it's different.
• ✅ Try using another embedding model like llama3-embed if it's available in your Ollama setup, which might be more compatible with MemOS.
• ✅ Check if MemOS allows configuring expected embedding dimensions — that could help align the inputs correctly.

Also, really nice work getting MemOS running on a custom port! If you’re planning to integrate it more tightly with Ollama (e.g. inference or embedding pipelines), double-check the API endpoint configuration and see if MemOS lets you customize how it connects to Ollama.