Update ollama.mdx

[hfnhf]

update one of two: add quotes around model name on

throws an error about base_url if the next line for model has no quotes around the model

update two of two:

change example model to a nice fully open model
https://ollama.com/library/[tulu3](https://ollama.com/library/tulu3)

end result:
no such error message about base_url (BAML 0.72.0 open-vsx extension is installed, screenshot to show)

---

Important

Fixes base_url error by quoting model name and updates example model to tulu3 in ollama.mdx.

• Behavior:
  • Adds quotes around the model name tulu3 in ollama.mdx to prevent a base_url error.
  • Changes example model from llama3 to tulu3 in ollama.mdx for a fully open model.

^{This description was created by for 07e0fd3. It will automatically update as commits are pushed.}

[vercel]

@hfnhf is attempting to deploy a commit to the Gloo Team on Vercel.

A member of the Team first needs to authorize it.

[ellipsis-dev]

👍 Looks good to me! Reviewed everything up to 07e0fd3 in 7 seconds

More details

• Looked at 13 lines of code in 1 files
• Skipped 0 files when reviewing.
• Skipped posting 1 drafted comments based on config settings.

1. fern/03-reference/baml/clients/providers/ollama.mdx:22

• Draft comment:
  The addition of quotes around the model name is correct and necessary to prevent syntax errors related to the base_url. This change ensures the model name is interpreted as a string.
• Reason this comment was not posted:
  Confidence changes required: 0%
  The PR correctly adds quotes around the model name, which is necessary for proper syntax in this context. The change aligns with the requirement to avoid errors related to base_url.

Workflow ID: wflow_JcT5e70memrIXu54

---

You can customize Ellipsis with 👍 / 👎 feedback, review rules, user-specific overrides, quiet mode, and more.

[hfnhf]

I'm still getting base_url error in the playground, even if the clients.baml doesn't show base_url error anymore, need to close while investigating what's actually going on

[aaronvg]

hmm this one works for me for example:

client<llm> Llama3 {
  // See https://docs.boundaryml.com/docs/snippets/clients/providers/ollama
  // to learn more about how to configure this client
  //
  // Note that you should run ollama using `OLLAMA_ORIGINS='*' ollama serve`
  // and you'll also need to `ollama pull llama3` to use this client
  provider openai-generic
  options {
    base_url "http://localhost:11434/v1"
    model "llama3.2:1b"
  }
}

[aaronvg]

your client in your function should say "CustomOllama"

[hfnhf]

Thank you aaronvg sir, I also was able to run tests

with resume.baml containing
...

function ExtractResume(resume: string) -> Resume {
  // Specify a client as provider/model-name
  // you can use custom LLM params with a custom client name from clients.baml like "client CustomHaiku"
  **client "CustomOllama"** 
  prompt #"
    
    {{ ctx.output_format }}
    Extract from this content:
    {{ resume }}

  "#
}

...

and clients.baml containing

client<llm> CustomOllama {
  provider "ollama"
  options {
    base_url "http://localhost:11434/v1"
    model "3thtulu3"
  }
}

I will wait before myself submitting any other pull request until after myself understanding the full context of the rest of the documentation.