feat(transcription): enable Vulkan GPU acceleration for Whisper inference

crates/transcription/Cargo.toml

@@ -8,7 +8,7 @@ description = "Speech-to-text engine wrappers, model management, and inference c
 kon-core = { path = "../core" }
 
 # Unified STT engine (Parakeet via ONNX, Whisper via whisper.cpp)
-transcribe-rs = { version = "0.3", features = ["onnx", "whisper-cpp"] }
+transcribe-rs = { version = "0.3", features = ["onnx", "whisper-cpp", "whisper-vulkan"] }
 
 # Async runtime for spawn_blocking
 tokio = { version = "1", features = ["rt", "sync"] }