enhancement: Add custom Jinja chat template option

Adds a new configuration option `chat_template` to the Llama.cpp extension, allowing users to define a custom Jinja chat template for the model.

The template can be provided via a new input field in the settings, and if set, it will be passed to the Llama.cpp backend using the `--chat-template` argument. This enhances flexibility for users who require specific chat formatting beyond the GGUF default.

The `chat_template` is added to the `LlamacppConfig` type and conditionally pushed to the command arguments if it's provided. The placeholder text provides an example of a Jinja template structure.

extensions/llamacpp-extension/settings.json

@@ -23,6 +23,18 @@
     "controllerType": "checkbox",
     "controllerProps": { "value": true }
   },
+  {
+      "key": "chat_template",
+      "title": "Custom Jinja Chat template",
+      "description": "Custom Jinja chat_template to be used for the model",
+      "controllerType": "input",
+      "controllerProps": {
+          "value": "",
+          "placeholder": "e.g., {% for message in messages %}...{% endfor %} (default is read from GGUF)",
+          "type": "text",
+          "textAlign": "right"
+      }
+  },
   {
     "key": "threads",
     "title": "Threads",

extensions/llamacpp-extension/src/index.ts

@@ -33,6 +33,7 @@ type LlamacppConfig = {
   version_backend: string
   auto_update_engine: boolean
   auto_unload: boolean
+  chat_template: string
   n_gpu_layers: number
   ctx_size: number
   threads: number
@@ -793,6 +794,7 @@ export default class llamacpp_extension extends AIEngine {
     }
 
     // Add remaining options from the interface
+    if (cfg.chat_template) args.push('--chat-template', cfg.chat_template)
     args.push('-ngl', String(cfg.n_gpu_layers > 0 ? cfg.n_gpu_layers : 100))
     if (cfg.threads > 0) args.push('--threads', String(cfg.threads))
     if (cfg.threads_batch > 0)