Codeseeker v2 does not always write the code

I am trying to use Codeseeker v2 on Ollama installed on my digitalocean, I did it and it works but sometimes it only writes the code on the chat

ALSO

I am trying to use the 236b version and it just doesn’t even show up as an option on Bolt

Hi,
can you please give more details.

What Model exactly did you run?
What are your hardware specs?
Did you set the context size in .env file?

At all this can happen, when the Model is not working very good together with bolt. You can try to tell it with “Write the code in the files”. Maybe it helps.

Hey, its okay I just changed to gemini, easier and cheaper, deepseek on ollama is too expensive to run on a gpu instance on vast.ai.

1 Like