Merge pull request #5962 from continuedev/nate/vllm-use-completions

Support completions with proxy

Author: tomasz-stefaniak

core/llm/llms/stubs/ContinueProxy.ts

@@ -85,7 +85,7 @@ class ContinueProxy extends OpenAI {
   }
 
   supportsCompletions(): boolean {
-    return false;
+    return true;
   }
 
   supportsFim(): boolean {

docs/docs/reference.md

@@ -550,8 +550,7 @@ models:
   - name: qwen2.5-coder-7b
     <<: *model_defaults
     model: qwen2.5-coder-7b
-    env:
-      useLegacyCompletionsEndpoint: false
+    useLegacyCompletionsEndpoint: false
     roles:
       - autocomplete
 ```

extensions/vscode/package-lock.json

@@ -2,7 +2,7 @@
   "name": "continue",
   "icon": "media/icon.png",
   "author": "Continue Dev, Inc",
-  "version": "1.1.41",
+  "version": "1.1.42",
   "repository": {
     "type": "git",
     "url": "https://github.com/continuedev/continue"