Repository navigation
Replies: 2 comments
|
There are currently no open-source LLMs that fit within 6 GB of memory and can work with HA-MCP. The memory required to store the context alone can easily exceed 6 GB. |
0 replies
|
That's funny, I actually have the exact same GPU on my spare computer that I keep running 24/7. I use the Qwen 4B VLM instruct on it for identifying images that my Frigate detects and for a few other little odds and ends. I tried using it with ha-mcp, it really didn't work well. I had to tweak a LOT of settings just to get it to see ha-mcp at all. It doesn't work very well but it can work with some of the lgihter tols if you have tool search mode on but I prsonally just use claude and chatgpt for my needs and it works great. |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Hello,
i am a bit new into the whole LLMs and mcp connectors, i am usign hermes with qwen3.5 4B Q4_K_M (gemini suggested that one for my rtx 2060 6gb)
i am getting a lot of errors while the llm is tryint to use the connector, is it becuase the model im usign isnt "smart" enough? or can i have some configuration errors?
i am getting things like the following:
All reactions