fix: PR review nits — response shape, landing copy, form reset

- /llm/generate's "model is downloading" branch was raising HTTPException(202, detail={...}), which wraps the payload in {"detail": ...} and forces clients to parse a success status as if it were an error. Switched to JSONResponse so the payload sits at the top level.
- The landing page's "Language Models" card advertised "Qwen 3.5" with sizes 4B/2B/0.8B; we ship Qwen3 at 0.6B/1.7B/4B. Aligned to what's actually in the binary.
- ProfileForm's discard-draft button reset the form without touching `personality` or `avatarFile`, so stale persona text and an attached avatar would survive the discard. The other three resets in the file already include both fields — this brings the discard path in line.

app/src/components/VoiceProfiles/ProfileForm.tsx

@@ -828,8 +828,10 @@ export function ProfileForm() {
                       name: '',
                       description: '',
                       language: 'en',
+                      personality: '',
                       sampleFile: undefined,
                       referenceText: '',
+                      avatarFile: undefined,
                     });
                     setSampleMode('record');
                   }}

backend/routes/llm.py

@@ -1,6 +1,7 @@
 """LLM inference endpoints."""
 
 from fastapi import APIRouter, HTTPException
+from fastapi.responses import JSONResponse
 
 from .. import models
 from ..backends import get_llm_model_configs
@@ -39,9 +40,9 @@ async def llm_generate(request: models.LLMGenerateRequest):
         task_manager.start_download(progress_model_name)
         create_background_task(download_llm_background())
 
-        raise HTTPException(
+        return JSONResponse(
             status_code=202,
-            detail={
+            content={
                 "message": f"Qwen3 {model_size} is being downloaded. Please wait and try again.",
                 "model_name": progress_model_name,
                 "downloading": True,

landing/src/components/SupportedModels.tsx

@@ -138,9 +138,9 @@ const MODEL_GROUPS: ModelGroup[] = [
     subtitle: 'Transcript refinement, persona replies, and on-device reasoning.',
     models: [
       {
-        name: 'Qwen 3.5',
+        name: 'Qwen3',
         author: 'Alibaba',
-        sizes: ['4B', '2B', '0.8B'],
+        sizes: ['4B', '1.7B', '0.6B'],
         description:
           'Powers transcript cleanup, persona voice replies, and the voice I/O loop. Shares its runtime with the TTS/STT stack — one model cache, one GPU story.',
         tags: [