Tag: speech recognition
-
Whisper.cpp vs Faster-Whisper: Why Speed Tests Lie
Lab benchmarks show whisper.cpp winning, but production flips the winner. Here's why speed tests miss memory pressure, cold starts, and streaming overhead.
-
Whisper Tiny vs faster-whisper: 3x Speed, 12% WER Gap
faster-whisper cuts Jetson Nano latency 68% but WER jumps to 20%. Real benchmarks, memory spikes, and the thermal throttling nobody mentions.
-
Whisper Architecture: How OpenAI’s Speech Model Works
Whisper's encoder-decoder uses 1.5GB VRAM for Large model. Break down the architecture and memory bottlenecks before mobile deployment.
-
Real-time Whisper Is a Battery Nightmare (Here’s How to Fix It)
Real-time Whisper drains 1% battery/min. VAD + adaptive inference + thermal throttling bring it down to 0.2%. Benchmarks on iPhone 13 Pro.