NVIDIA just dropped Nemotron-3.5-ASR: one 0.6B model, 40+ languages, streaming. parakeet.cpp already runs it. On a plain CPU, 2.5x faster than @NVIDIAAI 's Nemo runtime, output byte-for-byte…

An M5 Max with 40 GPU cores just beat an M3 Ultra with 80 on parakeet.cpp. 🔥 Every model. ~1.7× faster on average, up to ~2×. Half the cores. Repo: by super @mudler_it and it's already part of his…

parakeet.cpp now does batched transcription. Decode N clips in one pass and a single GB10 runs up to 12x faster at batch 16. Peak ~1,260 clips/s. CPU sees 3-5x. Same model, bit-for-bit identical…

用 C++ + ggml 重写 NVIDIA Parakeet ASR 的推理路径,不依赖 Python。 parakeet.cpp 把 NVIDIA NeMo Parakeet 系列 ASR 用 C++17 移植到 ggml,CPU 和 GPU 都能跑,推理时彻底不依赖 Python。

what a wonderful project: parakeet.cpp GGML based parakeet inference pipeline that's 2x faster than my ONNX parakeet pipeline on Apple Silicon! (Needed a few local patches to get it going)

saving you going through a painfully long engagement optimized thread, here's the GitHub:

More than just fast: - cache-aware streaming with real-time end-of-utterance - word-level timestamps and confidence - a flat C-API to drop it into anything