feat(multimodal): allow to template placeholders

[mudler]

Description

This PR allows to customize the placeholders for the images that are sent back to the model.

For context, some models/libraries have different way to express images, videos or audio placeholders. For example, llama.cpp backend expects images within an [img-ID] tag, but other backends/models (e.g. vLLM) use a different notation ( <|image_|>). This changeset allow to specify the notation in the template config section of the YAML file:

For example, to override defaults, now it is possible to set in the model configuration the following:

template:
  video: "<|video_{{.ID}}|> {{.Text}}"
  image: "<|image_{{.ID}}|> {{.Text}}"
  audio: "<|audio_{{.ID}}|> {{.Text}}"

Notes for Reviewers

This PR is related to #3670 #3602 #2318

Signed commits

• Yes, I signed my commits.

[netlify]

✅ Deploy Preview for localai ready!

Name	Link
🔨 Latest commit	03b3ce9
🔍 Latest deploy log	https://app.netlify.com/sites/localai/deploys/670006cb5994f80008c5e9ad
😎 Deploy Preview	https://deploy-preview-3728--localai.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify site configuration.