openbmb

/

MiniCPM-V-4_5-gguf

like 26

Follow

OpenBMB 1.86k

https://github.com/tc-mb/ollama/tree/MIniCPM-V
https://github.com/OpenSQZ/MiniCPM-V-CookBook/blob/main/deployment/ollama/minicpm-v4_5_ollama.md

Our PR has not been merged yet, you can use the branch we provided and deploy it using this document.

It cannot run on the Jackson series.

@redyuan43 What is the Jackson series device?
I have limited knowledge of it and may need more information or log files so I can help you.

not working on my Mac mini M4

Error: 500 Internal Server Error: llama runner process has terminated: error:attach failed: attach failed (Not allowed to attach to process.  Look in the console messages (Console.app), near the debugserver entries, when the attach failed.  The subsystem that denied the attach permission will likely have logged an informative message about why it was denied.)

It looks like you didn't start ./ollama serve

Same problem here, it's probably a problem with the ollama<->model connection, because all the rest of the models working just fine on my mac and I've also tried it with a Nvidia machine. So the problem is probably not a Mac specific.

@Yingzir @imagick
You can refer to my answer.

Below is a more detailed explanation:
Your error indicates that your code is making a network request but can't find the service.
Ollama requires a server to be enabled.

You need to run ./ollama serve in one command line.
Then, in another command line, you can run the model using ./ollama run xxx.

Hey tc-mb, the server is up and as I mentioned, it was tested with two different machines that succeeded with other models on the same exact environment.

@imagick
I think you're still using it incorrectly. Please refer to this document for guidance.
https://github.com/OpenSQZ/MiniCPM-V-CookBook/blob/main/deployment/ollama/minicpm-v4_5_ollama.md

https://ollama.com/openbmb/minicpm-v4.5
The ollama model repository is here, and usage is simple.
./ollama run openbmb/minicpm-v4.5

@tc-mb , thanks for the help, but I just made a reinstall and just a simple ollama run openbmb/minicpm-v4.5 and it ended up the same way, unlike other models which work just fine the same way.

Did you follow the documentation, clone my branch, compile ollama yourself (recommended), or use the official ollama?

@tc-mb thanks I didn't know I needed to use that branch so it finally worked, but that was the real reason and not the fact that I didn't use ollama serve

Ok, glad you've solved the problem.