fix: create WhisperState once at load time, reuse across all chunks

Previously create_state() was called for every 60s audio chunk, triggering
whisper_init_state() each time. This allocates ~700 MB of GPU compute buffers
(KV caches, CUDA workspace) and re-initialises the CUDA backend per chunk.

For a 101-minute audio (102 chunks), this caused 102 GPU re-initialisations
and VRAM allocation cycles. Under VRAM pressure from concurrent processes,
CUDA allocation failures occurred silently — whisper returned language
detection results but 0 segments.

Fix: create WhisperState once in Transcriber::load() and reuse it for every
transcription call. GPU memory is stable; no_context=true prevents KV-cache
contamination between chunks.

WhisperState is Send+Sync (explicitly declared in whisper-rs) and holds its
own Arc<WhisperInnerContext>, so the model weights stay alive even after
WhisperContext is dropped.

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

src/worker.rs

@@ -68,12 +68,17 @@ pub fn start(
 }
 
 /// Dedicated OS thread that owns the Transcriber (non-Send) and runs inference.
+///
+/// The Transcriber holds a single `WhisperState` that is reused for every chunk.
+/// GPU compute buffers (~700 MB) are allocated once at startup rather than on
+/// every call, eliminating per-chunk `whisper_init_state` overhead and the
+/// VRAM churn that caused intermittent 0-segment results.
 fn transcriber_thread(
     rx:         std::sync::mpsc::Receiver<TranscribeRequest>,
     model_path: PathBuf,
     gpu_device: u32,
 ) {
-    let transcriber = match Transcriber::load(&model_path, gpu_device) {
+    let mut transcriber = match Transcriber::load(&model_path, gpu_device) {
         Ok(t)  => t,
         Err(e) => {
             tracing::error!(error = %e, "failed to load whisper model — transcriber thread exiting");