feat: add whisper.cpp ROCm backend support for AMD GPU acceleration

- transcription.py: new _transcribe_remote_whispercpp() using /inference endpoint
- transcription.py: backend param routes to openai or whispercpp remote path
- config.py: whisper.backend default 'openai', alt 'whispercpp'
- pipeline.py: passes backend from config to transcribe_file
- settings: backend dropdown (OpenAI-compat / whisper.cpp)
- SETUP.md: whisper.cpp ROCm build and systemd setup instructions

whisper-cpp-server running on beastix :8080 (ROCm0, gfx1030, RX 6800 XT)

frontend/settings.js

@@ -53,6 +53,7 @@ async function loadConfig() {
   if (!r.ok) return;
   const cfg = await r.json();
   document.getElementById('audio-device').value = (cfg.audio && cfg.audio.device) || '';
+  document.getElementById('whisper-backend').value = (cfg.whisper && cfg.whisper.backend) || 'openai';
   document.getElementById('whisper-url').value = (cfg.whisper && cfg.whisper.base_url) || '';
   document.getElementById('whisper-model').value = (cfg.whisper && cfg.whisper.model) || 'large-v3';
   const ollamaUrl = (cfg.ollama && cfg.ollama.base_url) || 'http://localhost:11434';
@@ -96,6 +97,7 @@ document.getElementById('save-btn').addEventListener('click', async function() {
     whisper: {
       base_url: document.getElementById('whisper-url').value,
       model: document.getElementById('whisper-model').value,
+      backend: document.getElementById('whisper-backend').value,
     },
     ollama: {
       base_url: document.getElementById('ollama-url').value,